See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 62a15e681b5c2e4b_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/62a15e681b5c2e4b_train_data.json
  type:
    field_input: choices
    field_instruction: context
    field_output: question
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 4
eval_max_new_tokens: 128
eval_steps: 150
eval_table_size: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: Romain-XV/2a628107-27f0-4d52-8674-73dd1607b8cd
hub_repo: null
hub_strategy: checkpoint
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: 0.3
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
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 8280
micro_batch_size: 2
mlflow_experiment_name: /tmp/62a15e681b5c2e4b_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
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: 150
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: 5671d717-a2e8-4a40-a8e7-921bed3d2b08
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 5671d717-a2e8-4a40-a8e7-921bed3d2b08
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

2a628107-27f0-4d52-8674-73dd1607b8cd

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-Chat-v1.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1072

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 8280

Training results

Training Loss	Epoch	Step	Validation Loss
3.1019	0.0001	1	3.4293
0.416	0.0215	150	0.4823
0.3156	0.0430	300	0.3454
0.3356	0.0644	450	0.2731
0.1899	0.0859	600	0.2265
0.3036	0.1074	750	0.2347
0.2368	0.1289	900	0.2393
0.22	0.1504	1050	0.2257
0.1542	0.1718	1200	0.1879
0.1295	0.1933	1350	0.1783
0.4094	0.2148	1500	0.1729
0.1246	0.2363	1650	0.1742
0.2168	0.2578	1800	0.1713
0.1402	0.2792	1950	0.1603
0.1195	0.3007	2100	0.1463
0.1193	0.3222	2250	0.1481
0.1921	0.3437	2400	0.1379
0.0784	0.3652	2550	0.1308
0.1947	0.3867	2700	0.1271
0.1558	0.4081	2850	0.1301
0.1662	0.4296	3000	0.1299
0.0889	0.4511	3150	0.1255
0.0786	0.4726	3300	0.1066
0.1756	0.4941	3450	0.1142
0.0479	0.5155	3600	0.1162
0.1282	0.5370	3750	0.1085
0.0922	0.5585	3900	0.1072

Framework versions

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