See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: furiosa-ai/mlperf-gpt-j-6b
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 1c92e33d67cb4908_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/1c92e33d67cb4908_train_data.json
  type:
    field_instruction: input persona
    field_output: synthesized text
    format: '{instruction}'
    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: false
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: Alphatao/7e5f9625-ae50-4b29-96d9-2f7e443d22e2
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: 1284
micro_batch_size: 4
mlflow_experiment_name: /tmp/1c92e33d67cb4908_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: 1024
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: e51a9104-f9a0-4430-8a0c-2df2aa197b76
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: e51a9104-f9a0-4430-8a0c-2df2aa197b76
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

7e5f9625-ae50-4b29-96d9-2f7e443d22e2

This model is a fine-tuned version of furiosa-ai/mlperf-gpt-j-6b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9528

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: 1284

Training results

Training Loss	Epoch	Step	Validation Loss
17.0508	0.0007	1	2.2220
9.4732	0.0683	100	1.1640
8.7429	0.1366	200	1.1081
8.7836	0.2049	300	1.0705
8.2098	0.2732	400	1.0471
8.5538	0.3415	500	1.0258
7.9397	0.4098	600	1.0069
7.8745	0.4781	700	0.9919
7.7272	0.5464	800	0.9780
7.4231	0.6148	900	0.9677
8.2324	0.6831	1000	0.9599
7.8626	0.7514	1100	0.9545
7.7996	0.8197	1200	0.9528

Framework versions

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