tangled-0.4-0.5b-base

time python -B prepare_core_datasets.py

# 4096 x 4000
Progress: 100%|████████| 235/235 [1:24:19<00:00, 21.53s/it]
Workers are finished.██| 235/235 [1:24:19<00:00, 21.53s/it]

# 32768 x 500
Progress: 100%|████████| 235/235 [1:28:33<00:00, 22.61s/it]
Workers are finished.██| 235/235 [1:28:33<00:00, 22.61s/it]

i=0, block_size=4096, chunk_size=16384000, len(dataset)=2997148, len(dataset) * block_size=12276318208
Total number of tokens in the optimized dataset '../core-data-0-4096-4000' is 12276318208

i=1, block_size=32768, chunk_size=16384000, len(dataset)=374022, len(dataset) * block_size=12255952896
Total number of tokens in the optimized dataset '../core-data-1-32768-500' is 12255952896

real    172m58.892s
user    1150m0.879s
sys     11m49.907s

CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt pretrain --config pretrain-core-model.yaml

Seed set to 23
Time to instantiate model: 0.23 seconds.
Total parameters: 199,459,328
Verifying settings ...
Measured TFLOPs: 7069.07
Epoch 1 | iter 256 step 1 | loss train: 12.109, val: n/a | iter time: 378.87 ms (step) remaining time: 7 days, 8:11:36
Epoch 1 | iter 512 step 2 | loss train: 12.106, val: n/a | iter time: 346.80 ms (step) remaining time: 6 days, 12:20:04
Epoch 1 | iter 768 step 3 | loss train: 12.104, val: n/a | iter time: 348.23 ms (step) remaining time: 6 days, 5:58:13
Epoch 1 | iter 1024 step 4 | loss train: 12.093, val: n/a | iter time: 347.90 ms (step) remaining time: 6 days, 3:14:51
Epoch 1 | iter 1280 step 5 | loss train: 12.090, val: n/a | iter time: 348.40 ms (step) remaining time: 6 days, 1:15:13
Epoch 1 | iter 1536 step 6 | loss train: 12.070, val: n/a | iter time: 349.50 ms (step) remaining time: 5 days, 23:54:58
Epoch 1 | iter 1792 step 7 | loss train: 12.032, val: n/a | iter time: 348.92 ms (step) remaining time: 5 days, 22:57:52
Epoch 1 | iter 2048 step 8 | loss train: 12.009, val: n/a | iter time: 347.99 ms (step) remaining time: 5 days, 22:14:45
Epoch 1 | iter 2304 step 9 | loss train: 11.979, val: n/a | iter time: 348.45 ms (step) remaining time: 5 days, 21:40:34
Epoch 1 | iter 2560 step 10 | loss train: 11.940, val: n/a | iter time: 348.64 ms (step) remaining time: 5 days, 21:13:17
Epoch 1 | iter 2816 step 11 | loss train: 11.862, val: n/a | iter time: 347.73 ms (step) remaining time: 5 days, 20:50:38
# ...

Backup wandb:

mv wandb wandb-pretrain-core

Chat with model:

CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt chat ../out/pretrain-core/final

CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True time litgpt evaluate --tasks 'leaderboard' --out_dir '../evaluate/pretrain-core/leaderboard/' --batch_size 1 --dtype 'bfloat16' '../out/pretrain-core/final'

|                           Tasks                           |Version|Filter|n-shot|        Metric         |   |Value |   |Stderr|
|-----------------------------------------------------------|-------|------|-----:|-----------------------|---|-----:|---|------|
|leaderboard                                                |    N/A|      |      |                       |   |      |   |      |
| - leaderboard_bbh                                         |    N/A|      |      |                       |   |      |   |      |
|  - leaderboard_bbh_boolean_expressions                    |      1|none  |     3|acc_norm               |↑  |0.4600|±  |0.0316|
|  - leaderboard_bbh_causal_judgement                       |      1|none  |     3|acc_norm               |↑  |0.5027|±  |0.0367|
|  - leaderboard_bbh_date_understanding                     |      1|none  |     3|acc_norm               |↑  |0.1960|±  |0.0252|
|  - leaderboard_bbh_disambiguation_qa                      |      1|none  |     3|acc_norm               |↑  |0.3800|±  |0.0308|
|  - leaderboard_bbh_formal_fallacies                       |      1|none  |     3|acc_norm               |↑  |0.4680|±  |0.0316|
|  - leaderboard_bbh_geometric_shapes                       |      1|none  |     3|acc_norm               |↑  |0.0920|±  |0.0183|
|  - leaderboard_bbh_hyperbaton                             |      1|none  |     3|acc_norm               |↑  |0.5160|±  |0.0317|
|  - leaderboard_bbh_logical_deduction_five_objects         |      1|none  |     3|acc_norm               |↑  |0.1960|±  |0.0252|
|  - leaderboard_bbh_logical_deduction_seven_objects        |      1|none  |     3|acc_norm               |↑  |0.1520|±  |0.0228|
|  - leaderboard_bbh_logical_deduction_three_objects        |      1|none  |     3|acc_norm               |↑  |0.3440|±  |0.0301|
|  - leaderboard_bbh_movie_recommendation                   |      1|none  |     3|acc_norm               |↑  |0.2200|±  |0.0263|
|  - leaderboard_bbh_navigate                               |      1|none  |     3|acc_norm               |↑  |0.4200|±  |0.0313|
|  - leaderboard_bbh_object_counting                        |      1|none  |     3|acc_norm               |↑  |0.0560|±  |0.0146|
|  - leaderboard_bbh_penguins_in_a_table                    |      1|none  |     3|acc_norm               |↑  |0.2055|±  |0.0336|
|  - leaderboard_bbh_reasoning_about_colored_objects        |      1|none  |     3|acc_norm               |↑  |0.1520|±  |0.0228|
|  - leaderboard_bbh_ruin_names                             |      1|none  |     3|acc_norm               |↑  |0.2320|±  |0.0268|
|  - leaderboard_bbh_salient_translation_error_detection    |      1|none  |     3|acc_norm               |↑  |0.2200|±  |0.0263|
|  - leaderboard_bbh_snarks                                 |      1|none  |     3|acc_norm               |↑  |0.5393|±  |0.0375|
|  - leaderboard_bbh_sports_understanding                   |      1|none  |     3|acc_norm               |↑  |0.4600|±  |0.0316|
|  - leaderboard_bbh_temporal_sequences                     |      1|none  |     3|acc_norm               |↑  |0.2000|±  |0.0253|
|  - leaderboard_bbh_tracking_shuffled_objects_five_objects |      1|none  |     3|acc_norm               |↑  |0.2160|±  |0.0261|
|  - leaderboard_bbh_tracking_shuffled_objects_seven_objects|      1|none  |     3|acc_norm               |↑  |0.1480|±  |0.0225|
|  - leaderboard_bbh_tracking_shuffled_objects_three_objects|      1|none  |     3|acc_norm               |↑  |0.3520|±  |0.0303|
|  - leaderboard_bbh_web_of_lies                            |      1|none  |     3|acc_norm               |↑  |0.4880|±  |0.0317|
| - leaderboard_gpqa                                        |    N/A|      |      |                       |   |      |   |      |
|  - leaderboard_gpqa_diamond                               |      1|none  |     0|acc_norm               |↑  |0.2778|±  |0.0319|
|  - leaderboard_gpqa_extended                              |      1|none  |     0|acc_norm               |↑  |0.2509|±  |0.0186|
|  - leaderboard_gpqa_main                                  |      1|none  |     0|acc_norm               |↑  |0.2545|±  |0.0206|
| - leaderboard_ifeval                                      |      3|none  |     0|inst_level_loose_acc   |↑  |0.2026|±  |   N/A|
|                                                           |       |none  |     0|inst_level_strict_acc  |↑  |0.1679|±  |   N/A|
|                                                           |       |none  |     0|prompt_level_loose_acc |↑  |0.0869|±  |0.0121|
|                                                           |       |none  |     0|prompt_level_strict_acc|↑  |0.0684|±  |0.0109|
| - leaderboard_math_hard                                   |    N/A|      |      |                       |   |      |   |      |
|  - leaderboard_math_algebra_hard                          |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_counting_and_prob_hard                |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_geometry_hard                         |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_intermediate_algebra_hard             |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_num_theory_hard                       |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_prealgebra_hard                       |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
|  - leaderboard_math_precalculus_hard                      |      2|none  |     4|exact_match            |↑  |0.0000|±  |     0|
| - leaderboard_mmlu_pro                                    |    0.1|none  |     5|acc                    |↑  |0.1123|±  |0.0029|
| - leaderboard_musr                                        |    N/A|      |      |                       |   |      |   |      |
|  - leaderboard_musr_murder_mysteries                      |      1|none  |     0|acc_norm               |↑  |0.5000|±  |0.0317|
|  - leaderboard_musr_object_placements                     |      1|none  |     0|acc_norm               |↑  |0.2188|±  |0.0259|
|  - leaderboard_musr_team_allocation                       |      1|none  |     0|acc_norm               |↑  |0.3000|±  |0.0290|

3840.17user 10.07system 1:04:53elapsed 98%CPU (0avgtext+0avgdata 5976236maxresident)k
96664inputs+3177472outputs (14major+2239507minor)pagefaults 0swaps

litgpt convert_pretrained_checkpoint ../out/pretrain-core/final ../out/pretrain-core-converted
litgpt convert_from_litgpt ../out/pretrain-core-converted/ ../out/pretrain-core-converted

cp ../evaluate/pretrain-core/leaderboard/pytorch_model.bin ../out/pretrain-core-converted
mergekit-yaml merge-core-into-base.yaml ../out/pretrain-base-converted --clone-tensors

litgpt convert_to_litgpt --model_name "Qwen2.5-0.5B" --dtype bfloat16 ../out/pretrain-base-converted/
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt chat ../out/pretrain-base-converted/
cp -r ../out/pretrain-base-converted/ ../out/pretrain-base
rm ../out/pretrain-base/lit_model.pth ../out/pretrain-base/mergekit_config.yml ../out/pretrain-base/model_config.yaml