See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: false
base_model: Qwen/Qwen2.5-0.5B
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - 444b10db3e9b5a95_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/444b10db3e9b5a95_train_data.json
  type:
    field_instruction: link
    field_output: text
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
early_stopping_threshold: 1.0e-05
eval_max_new_tokens: 128
eval_steps: 150
eval_strategy: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: false
group_by_length: false
hub_model_id: clarxus/9cc234e5-b228-432d-9d1e-eb4026b6726b
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0004
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 150
lora_alpha: 16
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_memory:
  0: 75GB
max_steps: 500
micro_batch_size: 16
mlflow_experiment_name: /tmp/444b10db3e9b5a95_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 100
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
saves_per_epoch: 0
sequence_len: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.15
wandb_entity: techspear-hub
wandb_mode: online
wandb_name: 476fc048-ce25-43c4-90b8-98c02548b87c
wandb_project: Gradients-On-Seven
wandb_run: your_name
wandb_runid: 476fc048-ce25-43c4-90b8-98c02548b87c
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

9cc234e5-b228-432d-9d1e-eb4026b6726b

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.6482

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: 0.0004
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• 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: 100
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0006	1	2.8780
2.7789	0.0837	150	2.7107
2.691	0.1674	300	2.6640
2.6588	0.2510	450	2.6482

Framework versions

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