See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: false
base_model: NousResearch/Nous-Capybara-7B-V1
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - 2c82450a3b7fea1e_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/2c82450a3b7fea1e_train_data.json
  type:
    field_instruction: prompt_name
    field_output: original_prompt_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: 330
eval_strategy: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: false
group_by_length: false
hub_model_id: mrferr3t/d1d85a99-7e1f-4299-a206-90b2a876c08c
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: 330
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 
micro_batch_size: 4
mlflow_experiment_name: /tmp/2c82450a3b7fea1e_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: /workspace/hub_repo/last-checkpoint
s2_attention: null
sample_packing: false
save_steps: 330
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: 0.05
wandb_entity: null
wandb_mode: 
wandb_name: 139ec53e-2823-4cd9-bbc4-f90ecd22a3af
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 139ec53e-2823-4cd9-bbc4-f90ecd22a3af
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

d1d85a99-7e1f-4299-a206-90b2a876c08c

This model is a fine-tuned version of NousResearch/Nous-Capybara-7B-V1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0200

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0008	1	2.7393
0.5798	0.2573	324	0.0618
0.0462	0.5147	648	0.0341
0.0292	0.7720	972	0.0296
0.0318	1.0294	1296	0.0246
0.0334	1.2867	1620	0.0421
0.0291	1.5441	1944	0.0578
0.0326	1.8014	2268	0.0195
0.03	2.0588	2592	0.0325
0.0346	2.3161	2916	0.0223
0.023	2.5735	3240	0.0200

Framework versions

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