See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Llama-3.2-3B-Instruct
bf16: auto
chat_template: llama3
cosine_min_lr_ratio: 0.1
data_processes: 16
dataset_prepared_path: null
datasets:
- format: custom
  path: lavita/ChatDoctor-HealthCareMagic-100k
  type:
    field_input: input
    field_instruction: instruction
    field_output: output
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
device_map: '{'''':torch.cuda.current_device()}'
do_eval: true
early_stopping_patience: 1
eval_batch_size: 1
eval_sample_packing: false
eval_steps: 25
evaluation_strategy: steps
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 32
gradient_checkpointing: true
group_by_length: true
hub_model_id: cwaud/ea258833f53-4c8a-4b1a-9abf-59da4fa11e18
hub_repo: stevemonite
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: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lora_target_modules:
- q_proj
- v_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_memory:
  0: 70GiB
  1: 70GiB
  2: 70GiB
  3: 70GiB
max_steps: 888
micro_batch_size: 1
mlflow_experiment_name: lavita/ChatDoctor-HealthCareMagic-100k
model_type: AutoModelForCausalLM
num_epochs: 3
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.95
  adam_epsilon: 1e-5
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 50
save_strategy: steps
sequence_len: 2048
strict: false
tf32: false
tokenizer_type: AutoTokenizer
torch_compile: false
train_on_inputs: false
trust_remote_code: true
val_set_size: 50
wandb_entity: null
wandb_mode: online
wandb_project: Public_TuningSN
wandb_run: miner_id_24
wandb_runid: 153c7d6f
warmup_raio: 0.03
warmup_ratio: 0.04
weight_decay: 0.01
xformers_attention: null

ea258833f53-4c8a-4b1a-9abf-59da4fa11e18

This model is a fine-tuned version of unsloth/Llama-3.2-3B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.1079

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 35
• training_steps: 888

Training results

Training Loss	Epoch	Step	Validation Loss
3.0904	0.0003	1	3.2887
2.6898	0.0071	25	2.7006
2.5132	0.0143	50	2.5567
2.4234	0.0214	75	2.4513
2.5322	0.0285	100	2.4194
2.2951	0.0357	125	2.3581
2.331	0.0428	150	2.3471
2.2978	0.0499	175	2.3032
2.1733	0.0571	200	2.2889
2.194	0.0642	225	2.2525
2.3849	0.0714	250	2.2398
2.0697	0.0785	275	2.2127
2.5496	0.0856	300	2.2259
2.0852	0.0928	325	2.1999
2.2164	0.0999	350	2.2020
2.2373	0.1070	375	2.1708
2.2789	0.1142	400	2.1725
2.1254	0.1213	425	2.1471
2.11	0.1284	450	2.1469
2.0535	0.1356	475	2.1419
2.1039	0.1427	500	2.1362
1.9734	0.1498	525	2.1290
2.0061	0.1570	550	2.1158
2.1663	0.1641	575	2.1151
2.1725	0.1713	600	2.1112
2.2051	0.1784	625	2.1108
2.0556	0.1855	650	2.1073
1.9651	0.1927	675	2.1079

Framework versions

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