See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: facebook/opt-1.3b
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 5b1d7d2de5370ca4_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/5b1d7d2de5370ca4_train_data.json
  type:
    field_input: context
    field_instruction: question
    field_output: answers
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
gradient_clipping: 1.0
group_by_length: false
hub_model_id: dixedus/0f9303d1-8b35-410b-b787-8972b8a9deb0
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 5.0e-05
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 3
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: cosine
max_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/5b1d7d2de5370ca4_train_data.json
model_type: AutoModelForCausalLM
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
saves_per_epoch: 4
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: techspear-hub
wandb_mode: online
wandb_name: 46dd44b3-e4e4-4c44-8605-e8e8b6dd956e
wandb_project: Gradients-On-Eight
wandb_run: your_name
wandb_runid: 46dd44b3-e4e4-4c44-8605-e8e8b6dd956e
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

0f9303d1-8b35-410b-b787-8972b8a9deb0

This model is a fine-tuned version of facebook/opt-1.3b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5795

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0003	1	3.9752
14.4665	0.0031	9	3.3582
7.9998	0.0062	18	1.5246
4.7456	0.0093	27	1.0476
2.9898	0.0124	36	0.7092
2.7803	0.0155	45	0.6443
2.7031	0.0186	54	0.6148
2.3525	0.0217	63	0.5956
2.4174	0.0248	72	0.5864
2.4438	0.0278	81	0.5818
2.2881	0.0309	90	0.5798
2.2066	0.0340	99	0.5795

Framework versions

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