See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Meta-Llama-3.1-8B
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - thinker_train_data.json
  ds_type: json
  path: /workspace/input_data/thinker_train_data.json
  type:
    field_input: assistant
    field_instruction: reasoning
    field_output: user
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 10
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: besimray/miner_id_3_356953bd-f938-4862-a3a5-21d61fce48ce_1729861973
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: true
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: 500
micro_batch_size: 5
mlflow_experiment_name: /tmp/thinker_train_data.json
model_type: LlamaForCausalLM
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: 10
save_strategy: steps
sequence_len: 4096
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
val_set_size: 0.05
wandb_entity: besimray24-rayon
wandb_mode: online
wandb_project: Public_TuningSN
wandb_run: miner_id_24
wandb_runid: 356953bd-f938-4862-a3a5-21d61fce48ce
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

miner_id_3_356953bd-f938-4862-a3a5-21d61fce48ce_1729861973

This model is a fine-tuned version of unsloth/Meta-Llama-3.1-8B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7815

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: 5
• eval_batch_size: 5
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 20
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
1.5337	0.0056	1	1.4862
0.9214	0.0563	10	0.9408
0.835	0.1125	20	0.8533
0.7869	0.1688	30	0.8354
0.8884	0.2250	40	0.8179
0.774	0.2813	50	0.8094
0.8592	0.3376	60	0.8055
0.7419	0.3938	70	0.8004
0.7387	0.4501	80	0.7927
0.7656	0.5063	90	0.7874
0.7726	0.5626	100	0.7867
0.9268	0.6188	110	0.7775
0.8375	0.6751	120	0.7803
0.8536	0.7314	130	0.7765
0.6834	0.7876	140	0.7728
0.8245	0.8439	150	0.7661
0.6808	0.9001	160	0.7710
0.773	0.9564	170	0.7659
0.6604	1.0127	180	0.7712
0.5496	1.0689	190	0.7819
0.5153	1.1252	200	0.7815

Framework versions

• PEFT 0.13.2
• Transformers 4.45.2
• Pytorch 2.3.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1