See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: true
base_model: unsloth/tinyllama-chat
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - 6f716a15529b2fe2_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/6f716a15529b2fe2_train_data.json
  type:
    field_input: proposition
    field_instruction: instruction
    field_output: response
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
early_stopping_threshold: 0.0001
eval_max_new_tokens: 128
eval_steps: 114
eval_strategy: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: false
group_by_length: false
hub_model_id: mrferr3t/b2d8d09c-2fec-4b71-97a6-bc2a5685c8f7
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0004
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 114
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 
micro_batch_size: 32
mlflow_experiment_name: /tmp/6f716a15529b2fe2_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 100
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: 
s2_attention: null
sample_packing: false
save_steps: 114
saves_per_epoch: 0
sequence_len: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: 
wandb_name: 4ff326cf-471d-4d59-ae36-92a16644d352
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 4ff326cf-471d-4d59-ae36-92a16644d352
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

b2d8d09c-2fec-4b71-97a6-bc2a5685c8f7

This model is a fine-tuned version of unsloth/tinyllama-chat on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7085

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0004
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0004	1	1.2216
1.0283	0.0476	114	0.9294
0.8867	0.0953	228	0.8753
0.8604	0.1429	342	0.8479
0.8315	0.1905	456	0.8264
0.8176	0.2381	570	0.8114
0.7998	0.2858	684	0.7996
0.7906	0.3334	798	0.7886
0.7806	0.3810	912	0.7824
0.777	0.4287	1026	0.7745
0.7707	0.4763	1140	0.7686
0.7659	0.5239	1254	0.7642
0.7638	0.5715	1368	0.7571
0.7561	0.6192	1482	0.7543
0.7521	0.6668	1596	0.7510
0.7468	0.7144	1710	0.7476
0.7471	0.7621	1824	0.7433
0.7413	0.8097	1938	0.7407
0.7385	0.8573	2052	0.7358
0.7286	0.9050	2166	0.7352
0.7323	0.9526	2280	0.7323
0.7328	1.0002	2394	0.7303
0.6772	1.0478	2508	0.7296
0.6772	1.0955	2622	0.7296
0.6746	1.1431	2736	0.7292
0.682	1.1907	2850	0.7256
0.69	1.2384	2964	0.7252
0.6872	1.2860	3078	0.7244
0.6921	1.3336	3192	0.7218
0.6872	1.3812	3306	0.7193
0.686	1.4289	3420	0.7186
0.6857	1.4765	3534	0.7161
0.6834	1.5241	3648	0.7160
0.6831	1.5718	3762	0.7150
0.6892	1.6194	3876	0.7109
0.6778	1.6670	3990	0.7102
0.6772	1.7146	4104	0.7094
0.682	1.7623	4218	0.7086
0.681	1.8099	4332	0.7079
0.6807	1.8575	4446	0.7049
0.6839	1.9052	4560	0.7048
0.6821	1.9528	4674	0.7026
0.6858	2.0004	4788	0.7004
0.6241	2.0480	4902	0.7086
0.6297	2.0957	5016	0.7094
0.6282	2.1433	5130	0.7085

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1