Add files using upload-large-folder tool

.gitattributes

@@ -59,3 +59,21 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
 data_all_eng_slimpj/tokenizer234/tokenizer3/_text_document.idx filter=lfs diff=lfs merge=lfs -text
 data_all_eng_slimpj/tokenizer234/tokenizer4/_text_document.idx filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tokenizer234/tokenizer2/_text_document.idx filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalan filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalai filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalao filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalay filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbf filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalap filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbl filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalaq filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalav filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbg filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbd filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalaz filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbb filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalbc filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalaw filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalac filter=lfs diff=lfs merge=lfs -text
+data_all_eng_slimpj/tok_english-5/split40/finalau filter=lfs diff=lfs merge=lfs -text

data_all_eng_slimpj/shuffled/split4/finalaf

@@ -0,0 +1,2 @@
+{"text":"Radcliffe Camera \u00e4r en rund byggnad p\u00e5 Radcliffe Square i Oxford, England, ritad av arkitekten James Gibbs. Radcliffe Camera uppf\u00f6rdes 1737\u20131749 f\u00f6r att hysa Radcliffe Science Library, vilket flyttades till Oxford University Museum 1861. Byggnaden uppkallades efter l\u00e4karen John Radcliffe, som testamenterat sin egendom till Radcliffestiftelsen som finansierade uppf\u00f6randet av biblioteksbyggnaden.\n\nDen monumentala rotundaeffekten uppn\u00e5s genom en plan med tv\u00e5 koncentriska cylindrar och en rusticerad sockel. Den inre cylindern har en kupol. Det var ursprungligen en \u00f6ppen vestibul, men \u00f6ppningarna har nu glasats eller fyllts igen f\u00f6r att ge ytterligare en l\u00e4sesal.\n\nIdag tj\u00e4nar Radcliffe Camera som l\u00e4sesal till Bodleian Library.\n\nExterna l\u00e4nkar \n\nByggnader i Oxford\nOxfords universitet","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Das Schlagwort Wiege Europas bezeichnet das antike Griechenland, das man als kulturellen Ursprung Europas betrachtet hat. Diese Benennung gilt als antiquiert und wird besonders in begriffsbewussten Fachkreisen gemieden, da im Begriff der \"Wiege\" eine lineare, organische und teleologische Entwicklung der Geschichte impliziert ist und ein solches idealistisch fundiertes Geschichtsbild, nach dem eine Kultur eine Kindheitsphase habe, der eine Zeit der Bl\u00fcte und schlie\u00dflich der altersbedingte Verfall folge, als obsolet angesehen wird. Zudem bem\u00fchen sich die Kulturwissenschaften, ihre einst oft eurozentrische Sicht der Dinge zu \u00fcberwinden und der Beeinflussung der europ\u00e4ischen Kulturen auch durch au\u00dfereurop\u00e4ische Kulturkreise Rechnung zu tragen.\n\nSo gilt \"Europa\" heute als eine Kulturregion, die w\u00e4hrend ihrer Vorgeschichte aus verschiedenen Richtungen beeinflusst worden ist, so dass die Idee eines einzigen Ursprungs nicht mehr haltbar ist. Die vielf\u00e4ltigen Einfl\u00fcsse aus Asien und Afrika (etwa durch die Araber\/Mauren in S\u00fcditalien und Andalusien) und ihre Weiterentwicklung f\u00fchrten zur Diversit\u00e4t der Kulturen, wozu auch die Geomorphologie Europas beitrug.\n\nLiteratur \n Wilhelm Tielker: Der Mythos von der Idee Europa. Lit Verlag, 2003, ISBN 3-8258-6659-9.\n\nSiehe auch \n Wiege der Menschheit\n \u00c4gypten\n Zweistromland\n Knossos\n\nEurop\u00e4ische Geschichte","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}

data_all_eng_slimpj/tok_english-5/split40/finalac

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:056664dc7d309d4afd22badbe7f5ea1c0a56f5de3f48897837151a2d6a9a7b65
+size 9328031894

data_all_eng_slimpj/tok_english-5/split40/finalai

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1981fcad48e08f6cb11dbe493b24453fc592bc27b5634fbc9e511f4c8f86f69
+size 4525457209

data_all_eng_slimpj/tok_english-5/split40/finalan

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ead74829607320db6d3138835336e68828f31f718b20d2cf5fe72e81c43ae4cb
+size 4518498243

data_all_eng_slimpj/tok_english-5/split40/finalao

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc46539d2a2691c47103fda7f3b0314eb582a1e334d441dfc823d065edf1301f
+size 4524104827

data_all_eng_slimpj/tok_english-5/split40/finalap

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:807fa5c4a7ba384cfa4201aa0cfef6f71a7a4b26c5d0b239f6e551ee18504f85
+size 4528990575

data_all_eng_slimpj/tok_english-5/split40/finalaq

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d102d707fce6ba9e61befd02e42fb87b5e2a942f9c0b5572a321b0a682638bc
+size 4529436037

data_all_eng_slimpj/tok_english-5/split40/finalau

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb8bbcc35c41a5f322524aae66b111a6f3517d01ff5785ab386f5154636cbf35
+size 4523693816

data_all_eng_slimpj/tok_english-5/split40/finalav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72b98a30679da3a71cb9002de8978bbb4f9ea929ddd9a227b2fd4d56742c29b0
+size 4522500859

data_all_eng_slimpj/tok_english-5/split40/finalaw

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:841e8061777a3f337a29ba242bd22ba1f4aa7181770a3ccb60278c2c29949898
+size 4518334416

data_all_eng_slimpj/tok_english-5/split40/finalay

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e76549619d09272a888116e725c784c65ba02ca0fb8676103568432317b30f9
+size 7349379410

data_all_eng_slimpj/tok_english-5/split40/finalaz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e7e2b51455cadb60c1c4942602a0416596d0c8c7fbaf6bfa6d546ce0cdafe2d
+size 7342110019

data_all_eng_slimpj/tok_english-5/split40/finalbb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f65d5c48aed59dc7ba169b3c2f5f4a90226855255d8a0c064e73ee7e03572239
+size 7347117279

data_all_eng_slimpj/tok_english-5/split40/finalbc

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bf1953724bdb77e4c3f129cd96ba4cdc466b1699b737cfa8c7f6cd6a931ce54
+size 7320550269

data_all_eng_slimpj/tok_english-5/split40/finalbd

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:986d099b1fa216b75133acba31afe366946b3fc03a0883524610213208adaa64
+size 7314262453

data_all_eng_slimpj/tok_english-5/split40/finalbf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b4b7a5a9fd94e7783389cdfdf5b52126a396e7551ac9755365fe968eba9d42a
+size 7334996173

data_all_eng_slimpj/tok_english-5/split40/finalbg

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88331a0984d1c9f50819b5ae6cdbf50ce607ac33849bb628ed953c77b4b6b3df
+size 7362974588

data_all_eng_slimpj/tok_english-5/split40/finalbl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f2d6b6d46eefe83bb9b20c03229a038de5a499c3398f733cc4ae50dd603e218
+size 7333959488

data_all_eng_slimpj/tok_english-5/split40/finalbn

@@ -0,0 +1,14 @@
+{"text":"Vavilala is a village in Krishna of the Indian state of Andhra Pradesh. It is located in Tiruvuru mandal.\n\nReferences\n\nVillages in Krishna district","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"\u00ab\u041d\u0435\u043e\u043a\u043e\u0301\u043d\u0447\u0435\u043d\u043d\u0430\u044f \u043f\u043e\u0301\u0432\u0435\u0441\u0442\u044c\u00bb\u00a0\u2014 \u0441\u043e\u0432\u0435\u0442\u0441\u043a\u0438\u0439 \u0445\u0443\u0434\u043e\u0436\u0435\u0441\u0442\u0432\u0435\u043d\u043d\u044b\u0439 \u0444\u0438\u043b\u044c\u043c, \u043c\u0435\u043b\u043e\u0434\u0440\u0430\u043c\u0430, \u0441\u043d\u044f\u0442 \u0440\u0435\u0436\u0438\u0441\u0441\u0451\u0440\u043e\u043c \u0424\u0440\u0438\u0434\u0440\u0438\u0445\u043e\u043c \u042d\u0440\u043c\u043b\u0435\u0440\u043e\u043c, 1955 \u0433\u043e\u0434.\n\n\u041a\u0440\u0430\u0442\u043a\u043e\u0435 \u0441\u043e\u0434\u0435\u0440\u0436\u0430\u043d\u0438\u0435 \n\u0422\u0430\u043b\u0430\u043d\u0442\u043b\u0438\u0432\u043e\u0433\u043e \u043a\u043e\u0440\u0430\u0431\u043b\u0435\u0441\u0442\u0440\u043e\u0438\u0442\u0435\u043b\u044f \u042e\u0440\u0438\u044f \u0415\u0440\u0448\u043e\u0432\u0430 \u043d\u0430\u0441\u0442\u0438\u0433\u043b\u0430 \u0431\u0435\u0434\u0430\u00a0\u2014 \u043f\u043e\u0441\u043b\u0435 \u043d\u0435\u0441\u0447\u0430\u0441\u0442\u043d\u043e\u0433\u043e \u0441\u043b\u0443\u0447\u0430\u044f \u0435\u043c\u0443 \u043f\u0430\u0440\u0430\u043b\u0438\u0437\u043e\u0432\u0430\u043b\u043e \u043d\u043e\u0433\u0438 \u0438 \u043e\u043d \u043f\u0440\u0438\u043a\u043e\u0432\u0430\u043d \u043a \u043a\u0440\u043e\u0432\u0430\u0442\u0438. \u041d\u043e \u0415\u0440\u0448\u043e\u0432 \u043d\u0435 \u0441\u0434\u0430\u0451\u0442\u0441\u044f, \u043f\u0440\u043e\u0434\u043e\u043b\u0436\u0430\u0435\u0442 \u0430\u043a\u0442\u0438\u0432\u043d\u043e \u0436\u0438\u0442\u044c \u0438 \u0440\u0430\u0431\u043e\u0442\u0430\u0442\u044c. \u0418 \u043a\u0430\u0436\u0434\u043e\u0435 \u0443\u0442\u0440\u043e \u0441 \u043d\u0435\u0442\u0435\u0440\u043f\u0435\u043d\u0438\u0435\u043c \u0436\u0434\u0451\u0442 \u043f\u0440\u0438\u0445\u043e\u0434\u0430 \u0443\u0447\u0430\u0441\u0442\u043a\u043e\u0432\u043e\u0433\u043e \u0432\u0440\u0430\u0447\u0430 \u0415\u043b\u0438\u0437\u0430\u0432\u0435\u0442\u044b \u041c\u0443\u0440\u043e\u043c\u0446\u0435\u0432\u043e\u0439. \u0415\u043b\u0438\u0437\u0430\u0432\u0435\u0442\u0430 \u041c\u0430\u043a\u0441\u0438\u043c\u043e\u0432\u043d\u0430\u00a0\u2014 \u0432\u0434\u043e\u0432\u0430, \u043f\u043e\u043b\u043d\u043e\u0441\u0442\u044c\u044e \u043e\u0442\u0434\u0430\u0451\u0442 \u0441\u0435\u0431\u044f \u0441\u0432\u043e\u0435\u0439 \u043f\u0440\u043e\u0444\u0435\u0441\u0441\u0438\u0438. \u0421\u0438\u043b\u044c\u043d\u043e\u0435 \u0447\u0443\u0432\u0441\u0442\u0432\u043e, \u0432\u043e\u0437\u043d\u0438\u043a\u0448\u0435\u0435 \u043c\u0435\u0436\u0434\u0443 \u0431\u043e\u043b\u044c\u043d\u044b\u043c \u0438 \u0432\u0440\u0430\u0447\u043e\u043c,\u00a0\u2014 \u043b\u044e\u0431\u043e\u0432\u044c, \u043f\u043e\u043c\u043e\u0433\u0430\u0435\u0442 \u0441\u043e\u0442\u0432\u043e\u0440\u0438\u0442\u044c \u0447\u0443\u0434\u043e \u0438 \u0415\u0440\u0448\u043e\u0432 \u043d\u0430\u0447\u0438\u043d\u0430\u0435\u0442 \u0445\u043e\u0434\u0438\u0442\u044c.\n\n\u0412 \u0440\u043e\u043b\u044f\u0445 \n \u042d\u043b\u0438\u043d\u0430 \u0411\u044b\u0441\u0442\u0440\u0438\u0446\u043a\u0430\u044f\u00a0\u2014 \u0415\u043b\u0438\u0437\u0430\u0432\u0435\u0442\u0430 \u041c\u0430\u043a\u0441\u0438\u043c\u043e\u0432\u043d\u0430 \u041c\u0443\u0440\u043e\u043c\u0446\u0435\u0432\u0430, \u0443\u0447\u0430\u0441\u0442\u043a\u043e\u0432\u044b\u0439 \u0432\u0440\u0430\u0447\n \u0421\u0435\u0440\u0433\u0435\u0439 \u0411\u043e\u043d\u0434\u0430\u0440\u0447\u0443\u043a\u00a0\u2014 \u042e\u0440\u0438\u0439 \u0421\u0435\u0440\u0433\u0435\u0435\u0432\u0438\u0447 \u0415\u0440\u0448\u043e\u0432, \u0433\u043b\u0430\u0432\u043d\u044b\u0439 \u043a\u043e\u043d\u0441\u0442\u0440\u0443\u043a\u0442\u043e\u0440 \u0411\u0430\u043b\u0442\u0438\u0439\u0441\u043a\u043e\u0433\u043e \u0437\u0430\u0432\u043e\u0434\u0430\n \u0421\u043e\u0444\u044c\u044f \u0413\u0438\u0430\u0446\u0438\u043d\u0442\u043e\u0432\u0430\u00a0\u2014 \u0410\u043d\u043d\u0430 \u041a\u043e\u043d\u0441\u0442\u0430\u043d\u0442\u0438\u043d\u043e\u0432\u043d\u0430, \u043c\u0430\u0442\u044c \u0415\u0440\u0448\u043e\u0432\u0430\n \u0415\u0432\u0433\u0435\u043d\u0438\u0439 \u0421\u0430\u043c\u043e\u0439\u043b\u043e\u0432\u00a0\u2014 \u0410\u043b\u0435\u043a\u0441\u0430\u043d\u0434\u0440 \u0414\u0435\u043d\u0438\u0441\u043e\u0432\u0438\u0447 \u0410\u0433\u0430\u043d\u0438\u043d, \u0432\u0440\u0430\u0447-\u043d\u0435\u0432\u0440\u043e\u043f\u0430\u0442\u043e\u043b\u043e\u0433\n \u0415\u0432\u0433\u0435\u043d\u0438\u0439 \u041b\u0435\u0431\u0435\u0434\u0435\u0432\u00a0\u2014 \u0424\u0451\u0434\u043e\u0440 \u0418\u0432\u0430\u043d\u043e\u0432\u0438\u0447, \u0441\u0435\u043a\u0440\u0435\u0442\u0430\u0440\u044c \u0440\u0430\u0439\u043a\u043e\u043c\u0430\n \u0410\u043b\u0435\u043a\u0441\u0430\u043d\u0434\u0440 \u041b\u0430\u0440\u0438\u043a\u043e\u0432\u00a0\u2014 \u0421\u043f\u0438\u0440\u0438\u043d-\u0434\u0435\u0434\n \u0413\u0435\u0440\u043c\u0430\u043d \u0425\u043e\u0432\u0430\u043d\u043e\u0432\u00a0\u2014 \u0412\u0430\u0441\u0438\u043b\u0438\u0439 \u0421\u043f\u0438\u0440\u0438\u043d, \u043e\u0442\u0435\u0446 \u041d\u0430\u0441\u0442\u0435\u043d\u044c\u043a\u0438\n \u042e\u0440\u0438\u0439 \u0422\u043e\u043b\u0443\u0431\u0435\u0435\u0432\u00a0\u2014 \u041d\u0438\u043a\u043e\u043b\u0430\u0439 \u041d\u0438\u043a\u043e\u043b\u0430\u0435\u0432\u0438\u0447 \u0421\u043b\u0430\u0434\u043a\u043e\u0432, \u0432\u0435\u0434\u0443\u0449\u0438\u0439 \u0438\u043d\u0436\u0435\u043d\u0435\u0440\n \u042d\u0440\u0430\u0441\u0442 \u0413\u0430\u0440\u0438\u043d\u00a0\u2014 \u041a\u043e\u043b\u043e\u0441\u043a\u043e\u0432, \u0436\u0435\u043d\u0441\u043a\u0438\u0439 \u043f\u0430\u0440\u0438\u043a\u043c\u0430\u0445\u0435\u0440\n \u0410\u043d\u0442\u043e\u043d\u0438\u043d\u0430 \u0411\u043e\u0433\u0434\u0430\u043d\u043e\u0432\u0430\u00a0\u2014 \u0442\u0451\u0442\u044f \u041f\u043e\u043b\u044f\n \u0412\u043b\u0430\u0434\u0438\u043c\u0438\u0440 \u0412\u043e\u0440\u043e\u043d\u043e\u0432\u00a0\u2014 \u041f\u043e\u043d\u043e\u043c\u0430\u0440\u0451\u0432, \u043f\u0430\u0446\u0438\u0435\u043d\u0442\n \u0421\u0435\u0440\u0433\u0435\u0439 \u0414\u0432\u043e\u0440\u0435\u0446\u043a\u0438\u0439\u00a0\u2014 \u0421\u0435\u0440\u0433\u0435\u0439 \u0410\u043d\u0434\u0440\u0435\u0435\u0432, \u0441\u0442\u0443\u0434\u0435\u043d\u0442\n \u041e\u043b\u044c\u0433\u0430 \u0417\u0430\u0431\u043e\u0442\u043a\u0438\u043d\u0430\u00a0\u2014 \u041d\u0430\u0434\u044f, \u0441\u0442\u0443\u0434\u0435\u043d\u0442\u043a\u0430\n \u0421\u0435\u0440\u0433\u0435\u0439 \u041a\u0430\u0440\u043d\u043e\u0432\u0438\u0447-\u0412\u0430\u043b\u0443\u0430\u00a0\u2014 \u0412\u0430\u043b\u0435\u043d\u0442\u0438\u043d \u041e\u0441\u0438\u043f\u043e\u0432\u0438\u0447, \u043f\u0440\u043e\u0444\u0435\u0441\u0441\u043e\u0440\n \u0411\u043e\u0440\u0438\u0441 \u041b\u0451\u0441\u043a\u0438\u043d\u00a0\u2014 \u0441\u0438\u043c\u0443\u043b\u044f\u043d\u0442\n \u0412\u043b\u0430\u0434\u0438\u043c\u0438\u0440 \u0427\u043e\u0431\u0443\u0440\u00a0\u2014 \u0434\u0438\u0440\u0435\u043a\u0442\u043e\u0440 \u0437\u0430\u0432\u043e\u0434\u0430\n \u042e\u0440\u0438\u0439 \u0411\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u00a0\u2014 \u0433\u043b\u0430\u0432\u043d\u044b\u0439 \u0438\u043d\u0436\u0435\u043d\u0435\u0440 \u0437\u0430\u0432\u043e\u0434\u0430\n \u0413\u043b\u0435\u0431 \u0421\u0435\u043b\u044f\u043d\u0438\u043d\u00a0\u2014 \u0412\u043e\u043b\u043e\u0434\u044f, \u0441\u0442\u0443\u0434\u0435\u043d\u0442\n \u0412\u0430\u043b\u0435\u043d\u0442\u0438\u043d\u0430 \u0427\u0435\u043c\u0431\u0435\u0440\u0433\u00a0\u2014 \u043c\u0430\u0442\u044c \u0431\u043e\u043b\u044c\u043d\u043e\u0433\u043e\n \u0422\u0430\u043d\u044f \u0425\u043e\u0440\u0438\u0448\u043a\u043e\u00a0\u2014 \u041d\u0430\u0441\u0442\u0435\u043d\u044c\u043a\u0430 (\u043d\u0435\u0442 \u0432 \u0442\u0438\u0442\u0440\u0430\u0445)\n \u0410\u0440\u0438\u0430\u0434\u043d\u0430 \u041a\u0443\u0437\u043d\u0435\u0446\u043e\u0432\u0430\u00a0\u2014 \u0432\u044b\u043f\u0443\u0441\u043a\u043d\u0438\u0446\u0430 (\u043d\u0435\u0442 \u0432 \u0442\u0438\u0442\u0440\u0430\u0445)\n\n\u0421\u044a\u0451\u043c\u043e\u0447\u043d\u0430\u044f \u0433\u0440\u0443\u043f\u043f\u0430 \n \u0410\u0432\u0442\u043e\u0440 \u0441\u0446\u0435\u043d\u0430\u0440\u0438\u044f\u00a0\u2014 \u041a\u043e\u043d\u0441\u0442\u0430\u043d\u0442\u0438\u043d \u0418\u0441\u0430\u0435\u0432\n \u0420\u0435\u0436\u0438\u0441\u0441\u0451\u0440\u00a0\u2014 \u0424\u0440\u0438\u0434\u0440\u0438\u0445 \u042d\u0440\u043c\u043b\u0435\u0440\n \u041e\u043f\u0435\u0440\u0430\u0442\u043e\u0440\u00a0\u2014 \u0410\u043d\u0430\u0442\u043e\u043b\u0438\u0439 \u041d\u0430\u0437\u0430\u0440\u043e\u0432\n \u0425\u0443\u0434\u043e\u0436\u043d\u0438\u043a\u0438\u00a0\u2014 \u0411\u0435\u043b\u043b\u0430 \u041c\u0430\u043d\u0435\u0432\u0438\u0447, \u0418\u0441\u0430\u0430\u043a \u041a\u0430\u043f\u043b\u0430\u043d\n \u041a\u043e\u043c\u043f\u043e\u0437\u0438\u0442\u043e\u0440\u00a0\u2014 \u0413\u0430\u0432\u0440\u0438\u0438\u043b \u041f\u043e\u043f\u043e\u0432\n \u0414\u0438\u0440\u0435\u043a\u0442\u043e\u0440 \u043a\u0430\u0440\u0442\u0438\u043d\u044b\u00a0\u2014 \u0422\u0430\u043c\u0430\u0440\u0430 \u0421\u0430\u043c\u043e\u0437\u043d\u0430\u0435\u0432\u0430\n\n\u0414\u043e\u043f\u043e\u043b\u043d\u0438\u0442\u0435\u043b\u044c\u043d\u0430\u044f \u0438\u043d\u0444\u043e\u0440\u043c\u0430\u0446\u0438\u044f \n \u0412 \u0444\u0438\u043b\u044c\u043c\u0435 \u043f\u0440\u043e\u0437\u0432\u0443\u0447\u0430\u043b\u0430 \u043c\u0443\u0437\u044b\u043a\u0430 \u041f.\u00a0\u0418.\u00a0\u0427\u0430\u0439\u043a\u043e\u0432\u0441\u043a\u043e\u0433\u043e.\n\n\u0421\u0441\u044b\u043b\u043a\u0438 \n\n\u0424\u0438\u043b\u044c\u043c\u044b \u0421\u0421\u0421\u0420 1955 \u0433\u043e\u0434\u0430\n\u0424\u0438\u043b\u044c\u043c\u044b-\u043c\u0435\u043b\u043e\u0434\u0440\u0430\u043c\u044b \u0421\u0421\u0421\u0420\n\u0424\u0438\u043b\u044c\u043c\u044b \u0441\u0442\u0443\u0434\u0438\u0438 \u00ab\u041b\u0435\u043d\u0444\u0438\u043b\u044c\u043c\u00bb\n\u0424\u0438\u043b\u044c\u043c\u044b \u043e \u0432\u0440\u0430\u0447\u0430\u0445\n\u0424\u0438\u043b\u044c\u043c\u044b \u043e\u0431 \u0438\u043d\u0436\u0435\u043d\u0435\u0440\u0430\u0445\n\u0424\u0438\u043b\u044c\u043c\u044b \u043e\u0431 \u0438\u043d\u0432\u0430\u043b\u0438\u0434\u0430\u0445\n\u0424\u0438\u043b\u044c\u043c\u044b \u043e \u0421\u0430\u043d\u043a\u0442-\u041f\u0435\u0442\u0435\u0440\u0431\u0443\u0440\u0433\u0435\n\u0424\u0438\u043b\u044c\u043c\u044b \u0424\u0440\u0438\u0434\u0440\u0438\u0445\u0430 \u042d\u0440\u043c\u043b\u0435\u0440\u0430\n\u0424\u0438\u043b\u044c\u043c\u044b \u043f\u043e \u0441\u0446\u0435\u043d\u0430\u0440\u0438\u044f\u043c \u041a\u043e\u043d\u0441\u0442\u0430\u043d\u0442\u0438\u043d\u0430 \u0418\u0441\u0430\u0435\u0432\u0430","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"El Ejercito del Norte fue un ejercito conformado para combatir contra la Confederaci\u00f3n Per\u00fa-Boliviana.El 19 de mayo de 1837 el entonces encargado del manejo de las relaciones exteriores de la Confederaci\u00f3n Argentina, Juan Manuel de Rosas declar\u00f3 la guerra a la Confederaci\u00f3n Per\u00fa-Boliviana tanto por Tarija como por el hecho de que tropas peruanobolivianas invadieron la mayor parte de Jujuy, la Puna de Atacama y el norte de la provincia de Salta. Chile le hab\u00eda declarado tambi\u00e9n la guerra el 11 de noviembre de 1836, con el apoyo de peruanos contrarios a Santa Cruz y la Confederaci\u00f3n, entabl\u00e1ndose una alianza t\u00e1cita. Alejandro Heredia fue nombrado comandante del Ej\u00e9rcito del Norte (esto es, el ej\u00e9rcito argentino cuyas tropas estaban compuestas casi en su totalidad por biso\u00f1os soldados reclutas del noroeste).\n\nTal ej\u00e9rcito improvisado y mal pertrechado tuvo a Heredia como jefe m\u00e1ximo y a los generales Gregorio Paz y Manuel Virto como inmediatos comandantes. Pr\u00e1cticamente careci\u00f3 de todo apoyo log\u00edstico desde el resto de la Argentina (en parte porque las fuerzas argentinas de las otras regiones deb\u00edan afrontar otros conflictos).\n\nEl 20 de enero de 1839 las fuerzas restauradoras al mando del general chileno Manuel Bulnes y el peruano Ram\u00f3n Castilla logran la victoria de Yungay contra Santa Cruz que pone fin a la Confederaci\u00f3n Per\u00fa-Boliviana. El 14 de febrero de 1839, el nuevo presidente de Bolivia comunic\u00f3 el fin de la guerra y el 26 de abril de 1839 el gobierno argentino le puso fin oficialmente.\n\nHistoria \nEn 1837 estall\u00f3 la guerra entre la Confederaci\u00f3n Argentina y la Confederaci\u00f3n Per\u00fa-Boliviana, tanto por el conflicto por la provincia de Tarija, como por el hecho de que tropas peruanobolivianas invadieron la mayor parte de Jujuy, la Puna de Atacama y el norte de la provincia de Salta. Rosas nombr\u00f3 a Alejandro Heredia comandante del Ej\u00e9rcito del Norte (un ej\u00e9rcito argentino cuyas tropas estaban compuestas casi en su totalidad por biso\u00f1os soldados reclutas del Noroeste argentino) contra el \"tirano Santa Cruz\".\n\nSu plan era recuperar Tarija y Tupiza r\u00e1pidamente, y atacar Potos\u00ed en oto\u00f1o, a fin de que al enemigo le resultara dif\u00edcil avanzar durante el invierno. Sus oficiales principales fueron Gregorio Paz y Manuel Virtu. Pero su ej\u00e9rcito, improvisado y mal pertrechado careci\u00f3 de todo apoyo log\u00edstico desde el resto de la Argentina (principalmente porque las fuerzas argentinas de las otras regiones deb\u00edan afrontar otros graves conflictos \u2013 bloqueos y hostilidades franco-inglesas, ataques de los \"colorados\" y unitarios instalados en Montevideo, apoyados por Brasil, Francia e Inglaterra, y diversas tropas mercenarias). Tambi\u00e9n las luchas entre la poblaci\u00f3n gaucha de las provincias con las poblaciones de etnias ind\u00edgenas por ese entonces no integradas en la incipiente Argentina mantuvieron ocupadas a las provincias.\n\nLas operaciones militares al mando de Alejandro Heredia lograron, con muchas dificultades, liberar las zonas de Jujuy y de Salta que hab\u00edan sido invadidas. Pero no pudieron recuperar Tarija, al tener que enfrentar tropas mucho m\u00e1s numerosas, descansadas y mejor dirigidas por expertos oficiales mercenarios como el general alem\u00e1n Otto Philip Braun. Se lograron algunas victorias en posici\u00f3n defensiva, especialmente en Humahuaca, pero los avances fueron casi nulos.\n\nTras la ocupaci\u00f3n de casi todo Jujuy y el norte de Salta por parte de los peruano-bolivianos estos pudieron establecer un fuerte dispositivo ante las contraofensivas de las provincias argentinas: los caminos m\u00e1s directos como el de la Quebrada de Humahuaca o el de Iruya o el bastante m\u00e1s dif\u00edcil de la Puna de Atacama estaban bloqueados. Pero lo m\u00e1s grave quiz\u00e1s fue el paso del caudillo Eustaquio M\u00e9ndez al bando peruanoboliviano ya que en esa \u00e9poca la Confederaci\u00f3n Per\u00fa-boliviana parec\u00eda ofrecer un estado mucho m\u00e1s rico, pr\u00f3spero y estable que el de la Confederaci\u00f3n Argentina que no parec\u00eda poder salir de la feroz guerra civil que azot\u00f3 a la Argentina en esa \u00e9poca.\n\nAnte tantos obst\u00e1culos las tropas al mando de Heredia optaron por un rodeo de las zonas en donde era fuerte el enemigo, es decir un avance por el entonces casi desconocido Chaco hasta ingresar a Tarija, pero la bastante peque\u00f1a hueste argentina se encontr\u00f3 diezmada por las plagas que entonces presentaba la regi\u00f3n chaque\u00f1a (paludismo, tripanosomiasis, amebiasis) y un calor t\u00f3rrido extenuante. Las tropas argentinas del norte lograron reingresar a la regi\u00f3n de Tarija pero fueron esperadas por tropas peruanobolivianas mucho m\u00e1s numerosas, bien alimentadas, frescas y descansadas en la cuesta Montenegro\/Coyambuyo a poca distancia de la ciudad de Tarija, para agravar las contrariedades que sufrieron las biso\u00f1as tropas de la Confederaci\u00f3n Argentina en Tarija se encontraban refugiados muchos unitarios de origen argentino que no hesitaron en apoyar a los peruanobolivianos contra los federales argentinos.\n\nAunque la guerra no tuvo una conclusi\u00f3n clara, la derrota en Coyambuyo (llamada por los bolivianos y peruanos batalla de Montenegro, casi a las puertas de la ciudad de Tarija) signific\u00f3 un desprestigio para las tropas que lideraba Heredia y tambi\u00e9n para \u00e9l mismo. En la pr\u00e1ctica, termin\u00f3 en un empate, que ser\u00eda resuelto m\u00e1s tarde por el ej\u00e9rcito chileno en Yungay.\n\nPara entonces Hered\u00eda contaba con un poderoso ej\u00e9rcito. Sus milicias tucumanas en 1838 se compon\u00edan de un batall\u00f3n y diez regimientos de caballer\u00eda, cada uno formado por dos, tres o m\u00e1s escuadrones (de dos compa\u00f1\u00edas de 62 hombres). Estas fuerzas consum\u00edan hasta el 60% del las erogaciones provinciales, pero permitieron al caudillo ejercer su \u00abProtectorado\u00bb sobre Jujuy, Salta y Catamarca a trav\u00e9s de los 5000 hombres bajo su mando, veteranos de las campa\u00f1as ininterrumpidas que se ven\u00edan dando desde los inicios de su gobierno.\n\nReferencias","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Australische slijkekster\n Nieuwguinese slijkekster","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"\u00abThe Pop Hits\u00bb\u00a0\u2014 \u043f'\u044f\u0442\u0430 \u043a\u043e\u043c\u043f\u0456\u043b\u044f\u0446\u0456\u044f \u0448\u0432\u0435\u0434\u0441\u044c\u043a\u043e\u0433\u043e \u0433\u0443\u0440\u0442\u0443 \u00abRoxette\u00bb. \u0412\u0438\u043f\u0443\u0449\u0435\u043d\u0430 2003 \u0440\u043e\u043a\u0443.\n\n\u0421\u043f\u0438\u0441\u043e\u043a \u043a\u043e\u043c\u043f\u043e\u0437\u0438\u0446\u0456\u0439 \n \u00abOpportunity Nox\u00bb\n \u00abThe Look\u00bb\n \u00abDressed for Success\u00bb (US Single Mix)\n \u00abDangerous\u00bb (G.M. Remaster '03)\n \u00abJoyride\u00bb (Single edit)\n \u00abThe Big L\u00bb\n \u00abChurch of Your Heart\u00bb (G.M. Remaster '03)\n \u00abHow Do You Do!\u00bb\n \u00abSleeping in My Car\u00bb (Single edit)\n \u00abRun To You\u00bb (G.M. Remaster '03)\n \u00abJune Afternoon\u00bb\n \u00abStars\u00bb (G.M. Remaster '03)\n \u00abThe Centre of the Heart\u00bb (G.M. Remaster '03)\n \u00abReal Sugar\u00bb\n \u00abLittle Miss Sorrow\u00bb\n\n\u0411\u043e\u043d\u0443\u0441\u043d\u0438\u0439 EP \n \u00abStupid\u00bb\n \u00abMakin' Love To You\u00bb\n \u00abBetter Off On Her Own\u00bb\n \u00abBla Bla Bla Bla Bla (You Broke My Heart)\u00bb\n\n\u0421\u0438\u043d\u0433\u043b\u0438 \n \u00abOpportunity Nox\u00bb\n \u00abOpportunity Nox\u00bb\n \u00abFading Like a Flower (Everytime You Leave)\u00bb (Live)\n \u00abBreathe\u00bb (T&A Demo)\n\n\u041f\u043e\u0441\u0438\u043b\u0430\u043d\u043d\u044f \n \u041e\u0444\u0456\u0446\u0456\u0439\u043d\u0438\u0439 \u0441\u0430\u0439\u0442 \u0433\u0443\u0440\u0442\u0443 \n\n\u041c\u0443\u0437\u0438\u0447\u043d\u0456 \u0430\u043b\u044c\u0431\u043e\u043c\u0438 2003\n\u0417\u0431\u0456\u0440\u043a\u0438 Roxette\n\u0410\u043d\u0433\u043b\u043e\u043c\u043e\u0432\u043d\u0456 \u043c\u0443\u0437\u0438\u0447\u043d\u0456 \u0430\u043b\u044c\u0431\u043e\u043c\u0438","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Wej\u015bcie Carls\u00f3w. Plotkara \u2013 dalszy ci\u0105g serii ksi\u0105\u017cek m\u0142odzie\u017cowych Plotkara napisana przez Cecily von Ziegesar. W Polsce pierwszy tom zosta\u0142 wydany w 2008 roku przez wydawnictwo Amber.\n\nBohaterowie \nG\u0142\u00f3wni bohaterowie:\n Avery Carls \u2013 ma imi\u0119 po babci; jasnoz\u0142ociste w\u0142osy, kobaltowo-niebieskie oczy, goni\u0105ca za trendami, bogini jak z bajki w sukienkach od Marini; \"Toczy wojn\u0119\" z Jack Laurent; ona, Baby i Owen s\u0105 trojaczkami wychowanymi tylko przez matk\u0119; przeprowadzili si\u0119 do Nowego Jorku ze Sconset do Nantucket.\n Baby Carls  \u2013 by\u0142a nieoczekiwanym dzieckiem, najm\u0142odsza z tr\u00f3jki; t\u0119skni za swoim ch\u0142opakiem; na pocz\u0105tku nie rozstaje si\u0119 z jego bluz\u0105, kt\u00f3r\u0105 przez przypadek jej zostawi\u0142. Po jakim\u015b czasie spotyka J.P. i zakochuj\u0105 si\u0119 w sobie.\n Owen Carls  \u2013 brat Avery i Baby, najstarszy z trojaczk\u00f3w, ma imi\u0119 po dziadku; przystojniak o do\u015b\u0107 surowej urodzie, naga pier\u015b, z\u0142ociste w\u0142osy; nale\u017cy do dru\u017cyny p\u0142ywackiej w szkole \u015bw. Judy; nie mo\u017ce zapomnie\u0107 o dziewczynie, z kt\u00f3r\u0105 sp\u0119dzi\u0142 wyj\u0105tkow\u0105 noc. Przedstawi\u0142a si\u0119 jako Kat. Okazuje si\u0119, \u017ce tak naprawd\u0119 ma na imi\u0119 Kelsey i jest dziewczyn\u0105 jego nowego przyjaciela...\n Rhys Sterling  \u2013 kapitan dru\u017cyny p\u0142ywackiej w szkole \u015bw. Judy; nowy przyjaciel Owena. Marzy, \u017ce w ko\u0144cu zrobi to ze swoj\u0105 dziewczyn\u0105, kt\u00f3r\u0105 zna od ma\u0142ego, Kelsey; Jego matka prowadzi program pt. \"Herbatka z lady Sterling\".\n Jacqueline \"Jack\" Laurent  \u2013 zacz\u0119\u0142o si\u0119 od incydentu w sklepie. Jedna torebka, dwie dziewczyny. Jack i Avery si\u0119 nienawidz\u0105 i rywalizuj\u0105 ze sob\u0105. Chodzi\u0142a z J.P., ale zerwa\u0142 z ni\u0105, by by\u0107 z Baby Carls.\n Kelsey \"Kat\" Talmadge  \u2013 by\u0142a dziewczyna Rhysa. Znali si\u0119 od dziewi\u0105tego roku \u017cycia. Pozna\u0142a Owena zesz\u0142ego lata w Nantucket, gdzie oboje kochali si\u0119 po raz pierwszy na pla\u017cy i stracili dziewictwo. Da\u0142a mu bransoletk\u0119 z napisem KAT, by j\u0105 zapami\u0119ta\u0142 na zawsze. Po jakim\u015b czasie Kat odwiedzi\u0142a Rhysa w szkole i tak Kelsey i Owen ponownie si\u0119 spotkali. Kelsey rozsta\u0142a si\u0119 z Rhysem, a gdy ten odkry\u0142, \u017ce tajemniczym ch\u0142opakiem Kat jest Owen, pobili si\u0119.\n J.P. Cashman  \u2013 Jedyny syn Dicka i Tatiany Cashman. Chodzi do Riverside i od d\u0142u\u017cszego czasujest ch\u0142opakiem Jack. Jednak zostawi\u0142 j\u0105 dla Baby, kt\u00f3ra wyprowadza\u0142a jego psy. Po kr\u00f3tkim romansie z Baby, J.P. wraca do Jack\n\nKsi\u0105\u017cki \n Wej\u015bcie Carls\u00f3w. Plotkara 12 (The Caryles, 2008)\n Nigdy nie m\u00f3w do\u015b\u0107. Wej\u015bcie Carls\u00f3w. Plotkara 13 (You just can't get enough, 2008)\n Daj mi szans\u0119. Wej\u015bcie Carls\u00f3w. Plotkara 14 (Take a chance on me, 2009)\n Kochaj tylko tego, z kim jeste\u015b. Wej\u015bcie Carls\u00f3w. Plotkara 15 (Love the One You're With) \u2013 premiera ang: pa\u017adziernik 2009\n\nUwagi \n\nAmeryka\u0144skie utwory dla dzieci i m\u0142odzie\u017cy","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Sinistra Repubblicana \u2013 ex movimento politico italiano\nSinistra Repubblicana (Izquierda Republicana) \u2013 partito politico spagnolo","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Gwynn ap Nudd (Gwen fils de Nuz), est, dans la mythologie celtique galloise, un des souverains ou des messagers de l'Annwvyn, l'Autre Monde des Celtes.\n\nMythologie \n\nGwynn ap Nudd est le fils de Nudd et le fr\u00e8re d'Edeyrn (l'Yder fils de Nut de Chr\u00e9tien de Troyes). Son nom provient du celtique vindos qui signifie \u00ab blanc, loyal, \u00e9clatant \u00bb. Il est parfois assimil\u00e9 \u00e0 l'Irlandais Finn Mac Cumaill, dont le grand-p\u00e8re Nuada, roi des Tuatha D\u00e9 Danann, est l'\u00e9quivalent de Nudd.\n\nIl a un r\u00f4le psychopompe puisque l'une de ses fonctions est de guider les \u00e2mes des morts vers l'Annwvyn, accompagn\u00e9 d'une meute de chiens fantastiques.\n\nDans le conte arthurien Kulhwch et Olwen, Creiddylad, fille de Lludd Llaw Ereund, la plus belle fille de toutes les Bretagnes, s'enfuit avec Gwythyr, fils de Greidawl. Mais avant qu'ils aient pu consommer leur union, Gwynn enl\u00e8ve la jeune fille. L'amant l\u00e8ve une arm\u00e9e, mais le ravisseur est victorieux et fait prisonniers de nombreux guerriers. Ces nobles ne sont lib\u00e9r\u00e9s que sur l'intervention du roi Arthur, il d\u00e9cide en outre que Creiddylad demeurera chez son p\u00e8re, sans que les deux pr\u00e9tendants puissent l'approcher. Depuis cette histoire, tous les ans, aux calendes de mai, Gwynn et Gwythyr se battent pour Creiddylad, cela durera jusqu'au Jugement dernier. \u00c0 ce moment, le vainqueur du combat gagnera la demoiselle. Ce duel rituel symboliserait le passage de la saison sombre \u00e0 la saison claire, l'ann\u00e9e celtique \u00e9tait divis\u00e9e en deux.\n\nGwynn ap Nudd participe \u00e0 la chasse mythique d'Arthur contre le sanglier Twrch Trwyth.\n\nSources \n\n Les Quatre Branches du Mabinogi, conte Kulhwch et Olwen, traduit, pr\u00e9sent\u00e9 et annot\u00e9 par Pierre-Yves Lambert, Gallimard, coll. \u00ab L'aube des peuples \u00bb, Paris, 1993, .\n\nNote\n\nBibliographie \n\nDivinit\u00e9 celte\nMythologie celtique galloise\nPersonnage du cycle arthurien\nPsychopompe","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"KHYM is a Christian radio station broadcasting on 103.9 FM, licensed to Copeland, Kansas and serving the areas of Dodge City, Kansas, Garden City, Kansas, and Liberal, Kansas. The station is owned by Great Plains Christian Radio. KHYM began broadcasting December 23, 1997.\n\nThe station's format consists primarily of Christian music with midday and evening blocks of Christian talk and teaching.  Christian talk and teaching programs heard on KHYM include; In Touch with Dr. Charles Stanley, Back to the Bible, Leading The Way with Michael Youssef, Turning Point with David Jeremiah, Insight For Living with Chuck Swindoll, Focus on the Family.\n\nTranslators\nKHYM is also heard on translators throughout Kansas and Oklahoma, as well as on full power station KHEV 90.3 in Fairview, Oklahoma.\n\nKansas\n\nOklahoma\n\nReferences\n\nExternal links \nKHYM's website\n\nHYM\nRadio stations established in 1997\n1997 establishments in Kansas","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"O Parque Ariana, local onde se encontra o Museu Ariana foram uma oferta \u00e1 cidade de Genebra por Gustave Revilliod de la Rive, grande  coleccionador e mecenas e \u00faltimo descendente da fam\u00edlia \"'de la Rive'\", que o criou fez constru\u00eddo entre 1877 e 1884.\n\nO dom\u00ednio chamado Ariana ocupava uma superf\u00edcie de  45 hectares e encontra-se na margem direita do Lago Lemano em  Genebra. No testamento estava expressamente dito que \"o seu pav\u00e3o possa deambular livremente no parque\", e que \"o Senhor de La Rive a\u00ed seja enterrado e o parque seja aberto ao p\u00fablico\"\n\nPal\u00e1cio das Na\u00e7\u00f5es\nFoi no parque que se constru\u00edu o Pal\u00e1cio das Na\u00e7\u00f5es, sede da ONU em Genebra, organiza\u00e7\u00e3o que recebeu o prestigioso pr\u00e9mio de \"Certificat de r\u00e9serve naturelle\" atribu\u00eddo pela \"Fondation Nature et \u00c9conomie\" pela maneira como cuida da biodiversidade do parque Ariana. Na ponte que liga a antiga ala do Pal\u00e1cio com o nove edif\u00edcio, encontra-se o sarc\u00f3fago do Gustave Revilliod de la Rive, tal como os desejos testamentais o exigiam.\n\nMuseu\nO museu, que \u00e9 um dos mais antigos da cidade,  torna-se por testamento depois da morte de  \"de la Rive\" em 1890, propriedade cidade. Hoje  inclu\u00ed tamb\u00e9m o Academia Internacional da Cer\u00e2mica (AIC) - uma associa\u00e7\u00e3o fundada em 1952 - assim como o Museu Su\u00ed\u00e7o da cer\u00e2mica e do Vidreo.\n\nLiga\u00e7\u00f5es externas\n \n\nOfertas a Genebra\nParques e jardins de Genebra\nAtra\u00e7\u00f5es tur\u00edsticas de Genebra","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"The Lair of the White Worm \u2013 sz\u00f3sty album studyjny holenderskiej grupy muzycznej God Dethroned. Wydawnictwo ukaza\u0142o si\u0119 15 listopada 2004 roku nak\u0142adem wytw\u00f3rni muzycznej Metal Blade Records. W ramach promocji do pochodz\u0105cego z p\u0142yty utworu \"The Lair Of The White Worm\" zosta\u0142 zrealizowany teledysk.\n\nLista utwor\u00f3w \nOpracowano na podstawie materia\u0142u \u017ar\u00f3d\u0142owego.\n \"Nihilism\" \u2013 05:10\n \"Arch Enemy Spain\" \u2013 04:20\n \"Sigma Enigma\" \u2013 04:06\n \"The Lair of the White Worm\" \u2013 03:31\n \"Rusty Nails\" \u2013 06:04\n \"Loyal to the Crown of God Dethroned\" \u2013 03:25\n \"Last Zip of Spit\" \u2013 04:20\n \"The Grey Race\" \u2013 04:11\n \"Salt In Your Wounds\" \u2013 05:01\n\nPrzypisy \n\nAlbumy God Dethroned\nAlbumy Metal Blade Records\nAlbumy muzyczne wydane w roku 2004","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Los Premios Ondas son unos galardones entregados a los profesionales de radio, televisi\u00f3n, publicidad en radio y m\u00fasica. Son concedidos anualmente por Radio Barcelona, emisora de la SER del grupo PRISA, desde 1954. Son los primeros galardones de radio y televisi\u00f3n instituidos en Espa\u00f1a, por lo que gozan de gran prestigio.\n\nCategor\u00edas \nEn 2020 las categor\u00edas son las siguientes.\n\nRadio\n Mejor presentador o programa de radio hablada.\n Mejor programa o mejor tratamiento informativo de un acontecimiento.\n Premio a la trayectoria.\n Premio al mejor programa, radio o plataforma radiof\u00f3nica de emisi\u00f3n online.\n Premio al mejor presentador o programa de radio musical.\n\nPublicidad en radio\n Mejor campa\u00f1a de radio.\n Premio a la mejor agencia de radio.\n\nTelevisi\u00f3n\n Mejor programa de entretenimiento.\n Mejor programa de actualidad o cobertura especial.\n Mejor presentador.\n Mejor presentadora.\n Mejor serie espa\u00f1ola.\n Mejor int\u00e9rprete masculino en ficci\u00f3n\n Mejor int\u00e9rprete femenina en ficci\u00f3n\n Mejor programa emitido por emisoras o cadenas no nacionales.\n Mejor webserie de ficci\u00f3n o programa de emisi\u00f3n online.\n\nM\u00fasica\n Premio a la trayectoria.\n Premio al artista del a\u00f1o.\n Premio al espect\u00e1culo musical o concierto.\n\nPremios internacionales\n Mejores programas, empresas o profesionales de radio (2 premios).\n Mejores programas, empresas o profesionales de televisi\u00f3n (2 premios).\n\nHistoria \nLos Premios Ondas fueron creados por Manuel Tar\u00edn Iglesias, director de la emisora decana de la radio espa\u00f1ola, Radio Barcelona, coincidiendo con los actos de celebraci\u00f3n de su 30\u00ba aniversario. Su antecedente fue el Concurso Anual de Guiones, cuya primera edici\u00f3n tuvo lugar en 1953. El nombre del premio fue tomado de la desaparecida revista Ondas, editada tambi\u00e9n por Uni\u00f3n Radio, embri\u00f3n de la actual Cadena SER.\n\nLos Ondas se entregaron por primera vez en 1954, concretamente el 14 de noviembre, al ser esa la fecha de la primera emisi\u00f3n radiof\u00f3nica de Radio Barcelona. Inicialmente los Ondas distingu\u00edan solo a profesionales y programas de la radio espa\u00f1ola, siendo premiados en la primera edici\u00f3n Bobby Deglan\u00e9, Jos\u00e9 Iglesias \"El Zorro\" y Pedro Pablo Ayuso, entre otros.\n\nDos a\u00f1os m\u00e1s tarde, en 1956, crea la categor\u00eda de premios internacionales. En 1957 se incorpora la televisi\u00f3n, siendo Laura Valenzuela la primera galardonada. En 1971, el sacerdote sevillano Federico Mar\u00eda P\u00e9rez-Estudillo S\u00e1nchez gana un Premio Ondas por una misa flamenca cantada en Florencia.\n\nA finales de los a\u00f1os ochenta, la entrada del grupo Prisa en la propiedad de la Cadena SER supuso la ampliaci\u00f3n de los Ondas a nuevas categor\u00edas, el cine (categor\u00eda creada en 1991) y la m\u00fasica (1992). En 1999 se introdujo un reconocimiento a la publicidad radiof\u00f3nica.\n\nLa edici\u00f3n de 2009 tuvo su componente de pol\u00e9mica al serle concedido un premio (en el apartado de mejor presentador de televisi\u00f3n) a Jorge Javier V\u00e1zquez, lo que fue considerado \"vergonzoso\" por buena parte de los profesionales del sector, lo que llev\u00f3 incluso a que Carles Francino, que deb\u00eda entregarle el premio, se negara a hacerlo.\n\nEn el a\u00f1o 2017 los premios batieron su r\u00e9cord de candidaturas recibidas (que ya hab\u00eda establecido r\u00e9cord en 2015 y 2016) con m\u00e1s de 400 inscripciones en las diferentes categor\u00edas.\n\nLa idea de los https:\/\/premiosonda.com\/historias-de-los-premios-ondas se inicio con la idea de Ram\u00f3n Barbat, quien en 1953 se termin\u00f3 reuni\u00e9ndo con el consejo Directivo de la Cadena Ser.\n\nPalmar\u00e9s \nGanadores de las diferentes ediciones:\n\nReferencias\n\nEnlaces externos \n Sitio web oficial de los Premios Ondas\n\n \nPremios de periodismo de Espa\u00f1a","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Radcliffe Camera \u00e4r en rund byggnad p\u00e5 Radcliffe Square i Oxford, England, ritad av arkitekten James Gibbs. Radcliffe Camera uppf\u00f6rdes 1737\u20131749 f\u00f6r att hysa Radcliffe Science Library, vilket flyttades till Oxford University Museum 1861. Byggnaden uppkallades efter l\u00e4karen John Radcliffe, som testamenterat sin egendom till Radcliffestiftelsen som finansierade uppf\u00f6randet av biblioteksbyggnaden.\n\nDen monumentala rotundaeffekten uppn\u00e5s genom en plan med tv\u00e5 koncentriska cylindrar och en rusticerad sockel. Den inre cylindern har en kupol. Det var ursprungligen en \u00f6ppen vestibul, men \u00f6ppningarna har nu glasats eller fyllts igen f\u00f6r att ge ytterligare en l\u00e4sesal.\n\nIdag tj\u00e4nar Radcliffe Camera som l\u00e4sesal till Bodleian Library.\n\nExterna l\u00e4nkar \n\nByggnader i Oxford\nOxfords universitet","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}
+{"text":"Das Schlagwort Wiege Europas bezeichnet das antike Griechenland, das man als kulturellen Ursprung Europas betrachtet hat. Diese Benennung gilt als antiquiert und wird besonders in begriffsbewussten Fachkreisen gemieden, da im Begriff der \"Wiege\" eine lineare, organische und teleologische Entwicklung der Geschichte impliziert ist und ein solches idealistisch fundiertes Geschichtsbild, nach dem eine Kultur eine Kindheitsphase habe, der eine Zeit der Bl\u00fcte und schlie\u00dflich der altersbedingte Verfall folge, als obsolet angesehen wird. Zudem bem\u00fchen sich die Kulturwissenschaften, ihre einst oft eurozentrische Sicht der Dinge zu \u00fcberwinden und der Beeinflussung der europ\u00e4ischen Kulturen auch durch au\u00dfereurop\u00e4ische Kulturkreise Rechnung zu tragen.\n\nSo gilt \"Europa\" heute als eine Kulturregion, die w\u00e4hrend ihrer Vorgeschichte aus verschiedenen Richtungen beeinflusst worden ist, so dass die Idee eines einzigen Ursprungs nicht mehr haltbar ist. Die vielf\u00e4ltigen Einfl\u00fcsse aus Asien und Afrika (etwa durch die Araber\/Mauren in S\u00fcditalien und Andalusien) und ihre Weiterentwicklung f\u00fchrten zur Diversit\u00e4t der Kulturen, wozu auch die Geomorphologie Europas beitrug.\n\nLiteratur \n Wilhelm Tielker: Der Mythos von der Idee Europa. Lit Verlag, 2003, ISBN 3-8258-6659-9.\n\nSiehe auch \n Wiege der Menschheit\n \u00c4gypten\n Zweistromland\n Knossos\n\nEurop\u00e4ische Geschichte","meta":{"redpajama_set_name":"RedPajamaWikipedia"}}

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_11.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-11
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-11
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-11-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_11.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_18.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-18
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-18
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-18-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_18.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_23.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-23
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-23
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-23-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_23.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_24.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-24
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-24
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-24-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_24.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_27.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-27
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-27
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-27-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_27.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_28.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-28
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-28
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-28-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_28.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_3.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-3
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-3
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-3-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_3.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_32.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-32
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-32
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-32-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_32.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_35.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-35
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-35
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-35-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_35.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_6.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-6
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-6
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-6-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_6.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tok_english-5/tok_jobs/tokenizer_9.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-5-9
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-5-9
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-5-9-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/hineng/data_all_eng_slimpj/tok_english-5/tok_files/tokenizer_9.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

data_all_eng_slimpj/tokenizer234/tokenizer2/_text_document.idx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29726ee729e270286affd4074cb2c682577ac1ab9de4bedceec9f05b98ce1234
+size 2156209142

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-00.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef61742381043e3d60e1e570c66810cd0085411ccfbc41d4768ca0884807b15d
+size 6050967032

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-01.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2870cc6aa948102a8a7f2a87435fd7d9a86a3944d7e3d1ba6285ac7989977a58
+size 6050965940

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-02.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e57ce8a8496a97a64f8584ff027179201fffe5537359f76e00cb9b7ca954bfd5
+size 6050966316

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-03.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55162155c63fce5ab9ab53b4b04602abd607aa9c3cae06dc21d6487733628b4c
+size 6050966452

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-04.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4163d0383766f5c77d085ed293132461e8d83a2605bf8207dd3bc65100c0a99f
+size 6050966432

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-05.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf51f8819117a9a1adaf4b52360b2d92dbe00ab358afecb7c37441f11006feb1
+size 6050966413

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-06.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:728234b74eec774020d0212f92a9eb1439d63b4fa031b51b739d8d03429c40e8
+size 6050966407

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-07.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12d282cf490e871107a92c83b913664617e4c152b01bace7c0762c68ea2fae7d
+size 6050966472

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-08.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b56b2b8ae1e9e7e2e83d5178fa724aaaccfa1d45e393a59f4d36a4cd56944f19
+size 6050966372

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-09.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:225f3db6c1d70212798013c4113a3319d3ce43df3e81b07b1c4d84e716ffdbcc
+size 6050966536

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-10.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e667ef68b3af6a0f332b3eb91e4434474139ce12b567f93536a6632a91a964e
+size 6050966335

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-11.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21f303c4b2b25ec7d97b0dbf6fbbe80b4c087c476f50d7cd8aa70144d451c8cb
+size 6050966515

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-12.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1b71a8d6aa602e58b84df896fac8ac179b02e3cb3a1238b32d684ef97f5f557
+size 6050967087

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-13.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb1780b8d4df05d7f6b134795fe80b4ac07698722caa3527cdd31ac9500503fd
+size 6050965643

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-14.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c029f4f4b35fa9e612a580aea817723153627f6a8981b9b04ab9fd2d274b4b4c
+size 6050966645

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-15.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bb4c67d8aa0c0982aecef2b81b367f8242fca7f61d465b9b7e8028ec1b1416b
+size 6050966617

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-16.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7b4d7252680dde5758427c421d646a981a9c17f39579298dd13daed88b014e4
+size 6050966182

data_all_eng_slimpj/tokenizer234/tokenizer2/split__text_document.bin/part-17.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad2da141d65983e405cc360eab07aeabee68cc38548a16df0d5e2f1c9260ef11
+size 6050966798