See axolotl config

axolotl version: 0.4.1

adapter: qlora
base_model: OpenBuddy/openbuddy-llama2-13b-v8.1-fp16
bf16: auto
chat_template: llama3
dataloader_num_workers: 6
dataset_prepared_path: null
datasets:
- data_files:
  - 258997f5c768a462_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/258997f5c768a462_train_data.json
  type:
    field_instruction: abstractText
    field_output: meshMajor
    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: 50
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: true
group_by_length: true
hub_model_id: error577/b3b5f3c6-cda5-4eaf-958b-be1f5e6069f7
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0003
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.3
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: 300
micro_batch_size: 1
mlflow_experiment_name: /tmp/258997f5c768a462_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 50
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: 50
sequence_len: 256
special_tokens:
  pad_token: </s>
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: 6d543dd5-ac1c-4d90-9cb2-8669fccd9d2c
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 6d543dd5-ac1c-4d90-9cb2-8669fccd9d2c
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

b3b5f3c6-cda5-4eaf-958b-be1f5e6069f7

This model is a fine-tuned version of OpenBuddy/openbuddy-llama2-13b-v8.1-fp16 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4757

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: 0.0003
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• 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: 300

Training results

Training Loss	Epoch	Step	Validation Loss
0.9502	0.0003	1	1.3877
0.6039	0.0161	50	0.5690
0.5687	0.0323	100	0.5499
0.5451	0.0484	150	0.5127
0.6351	0.0646	200	0.4887
0.6928	0.0807	250	0.4805
0.6263	0.0969	300	0.4757

Framework versions

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