See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: trl-internal-testing/tiny-random-LlamaForCausalLM
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 1370a6dd48e2ecba_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/1370a6dd48e2ecba_train_data.json
  type:
    field_input: document_description
    field_instruction: document_type
    field_output: generated_text
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
device_map:
  ? ''
  : 0,1,2,3,4,5,6,7
early_stopping_patience: 2
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: true
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: Alphatao/6978920a-5ce2-4670-bbc6-1d4638fcde76
hub_repo: null
hub_strategy: null
hub_token: null
learning_rate: 0.0002
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 21483
micro_batch_size: 4
mlflow_experiment_name: /tmp/1370a6dd48e2ecba_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
sequence_len: 2048
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.04
wandb_entity: null
wandb_mode: online
wandb_name: 1ea730f0-95e7-41bf-88e5-468b7a961401
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 1ea730f0-95e7-41bf-88e5-468b7a961401
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

6978920a-5ce2-4670-bbc6-1d4638fcde76

This model is a fine-tuned version of trl-internal-testing/tiny-random-LlamaForCausalLM on the None dataset. It achieves the following results on the evaluation set:

• Loss: 10.3137

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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: 10
• training_steps: 16422

Training results

Training Loss	Epoch	Step	Validation Loss
10.3817	0.0006	1	10.3802
10.3412	0.0609	100	10.3412
10.3369	0.1218	200	10.3334
10.3296	0.1827	300	10.3301
10.3243	0.2436	400	10.3274
10.3286	0.3045	500	10.3255
10.3294	0.3654	600	10.3241
10.3261	0.4262	700	10.3229
10.3291	0.4871	800	10.3222
10.3279	0.5480	900	10.3213
10.3223	0.6089	1000	10.3205
10.3202	0.6698	1100	10.3198
10.3244	0.7307	1200	10.3192
10.3198	0.7916	1300	10.3189
10.3302	0.8525	1400	10.3182
10.3191	0.9134	1500	10.3178
10.3266	0.9743	1600	10.3172
10.2112	1.0352	1700	10.3173
10.3815	1.0961	1800	10.3169
9.7867	1.1569	1900	10.3164
10.2159	1.2178	2000	10.3163
10.4188	1.2787	2100	10.3158
10.2751	1.3396	2200	10.3156
10.0096	1.4005	2300	10.3157
10.4669	1.4614	2400	10.3152
10.3051	1.5223	2500	10.3153
9.9095	1.5832	2600	10.3148
10.847	1.6441	2700	10.3146
10.162	1.7050	2800	10.3146
10.6745	1.7659	2900	10.3144
10.2911	1.8268	3000	10.3142
10.3342	1.8877	3100	10.3141
10.7718	1.9485	3200	10.3140
10.625	2.0094	3300	10.3139
10.3683	2.0703	3400	10.3137
9.9419	2.1312	3500	10.3138
10.1016	2.1921	3600	10.3137

Framework versions

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