See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: llamafactory/tiny-random-Llama-3
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - c30561d2c881448a_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/c30561d2c881448a_train_data.json
  type:
    field_input: tools
    field_instruction: func_name
    field_output: func_desc
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
device_map:
  ? ''
  : 0,1,2,3,4,5,6,7
early_stopping_patience: 2
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: true
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: Alphatao/a683dec5-2cfe-462b-8878-a60b574dc35d
hub_repo: null
hub_strategy: null
hub_token: null
learning_rate: 0.0002
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 7680
micro_batch_size: 4
mlflow_experiment_name: /tmp/c30561d2c881448a_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
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: 100
sequence_len: 2048
special_tokens:
  pad_token: <|eot_id|>
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.04
wandb_entity: null
wandb_mode: online
wandb_name: 875a76e4-786a-481a-8e49-13a443cf2d23
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 875a76e4-786a-481a-8e49-13a443cf2d23
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

a683dec5-2cfe-462b-8878-a60b574dc35d

This model is a fine-tuned version of llamafactory/tiny-random-Llama-3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 11.6630

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• 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: 2396

Training results

Training Loss	Epoch	Step	Validation Loss
11.7563	0.0008	1	11.7576
11.7084	0.0835	100	11.7043
11.7003	0.1670	200	11.6870
11.6836	0.2504	300	11.6811
11.6833	0.3339	400	11.6764
11.6782	0.4174	500	11.6740
11.6975	0.5009	600	11.6715
11.685	0.5844	700	11.6700
11.6776	0.6678	800	11.6687
11.6638	0.7513	900	11.6674
11.6674	0.8348	1000	11.6667
11.6872	0.9183	1100	11.6660
10.9085	1.0023	1200	11.6655
12.0933	1.0858	1300	11.6651
12.0125	1.1693	1400	11.6643
11.8914	1.2527	1500	11.6642
12.0877	1.3362	1600	11.6638
10.9328	1.4197	1700	11.6634
11.7329	1.5032	1800	11.6634
12.6102	1.5867	1900	11.6633
12.1426	1.6701	2000	11.6631
10.1984	1.7536	2100	11.6630
11.7955	1.8371	2200	11.6630
11.144	1.9206	2300	11.6630

Framework versions

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