See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_resume_from_checkpoints: true
base_model: TinyLlama/TinyLlama-1.1B-Chat-v0.6
bf16: auto
chat_template: llama3
dataset_prepared_path: null
dataset_processes: 6
datasets:
- data_files:
  - a9d4324f1e33f9ae_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/a9d4324f1e33f9ae_train_data.json
  type:
    field_instruction: init_prompt
    field_output: init_response
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 200
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/e044dcfc-83e2-4fa2-aad2-875a52316c8d
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: null
micro_batch_size: 16
mlflow_experiment_name: /tmp/a9d4324f1e33f9ae_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: 200
sequence_len: 256
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.005
wandb_entity: null
wandb_mode: online
wandb_name: 2a9710ae-957a-477d-91f9-0a8bce2d110f
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 2a9710ae-957a-477d-91f9-0a8bce2d110f
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

e044dcfc-83e2-4fa2-aad2-875a52316c8d

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-Chat-v0.6 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2170

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• 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: 30
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.2866	0.0007	1	1.2665
0.7368	0.1473	200	0.7152
0.6216	0.2947	400	0.6414
0.5531	0.4420	600	0.5916
0.4998	0.5893	800	0.5468
0.4112	0.7366	1000	0.5108
0.4076	0.8840	1200	0.4649
0.2988	1.0313	1400	0.4229
0.3479	1.1786	1600	0.3908
0.3133	1.3260	1800	0.3472
0.2523	1.4733	2000	0.3261
0.2227	1.6206	2200	0.2928
0.2482	1.7680	2400	0.2750
0.2354	1.9153	2600	0.2585
0.1221	2.0626	2800	0.2463
0.1213	2.2099	3000	0.2326
0.132	2.3573	3200	0.2288
0.1477	2.5046	3400	0.2223
0.1168	2.6519	3600	0.2196
0.1501	2.7993	3800	0.2174
0.1196	2.9466	4000	0.2170

Framework versions

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