adapter: qlora
base_model: NousResearch/Yarn-Llama-2-7b-64k
bf16: auto
chat_template: llama3
dataloader_num_workers: 6
dataset_prepared_path: null
datasets:
- data_files:
  - 1dc9438022262def_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/1dc9438022262def_train_data.json
  type:
    field_input: Source
    field_instruction: Script
    field_output: Content
    format: '{instruction} {input}'
    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: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: true
group_by_length: true
hub_model_id: error577/6c5e4288-6fb4-4569-a5d1-eeb97cd6a2f5
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: 128
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 500
micro_batch_size: 1
mlflow_experiment_name: /tmp/1dc9438022262def_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: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.01
wandb_entity: null
wandb_mode: online
wandb_name: 4a4ff803-1f36-452e-a6bd-d415e53ba94f
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 4a4ff803-1f36-452e-a6bd-d415e53ba94f
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

6c5e4288-6fb4-4569-a5d1-eeb97cd6a2f5

This model is a fine-tuned version of NousResearch/Yarn-Llama-2-7b-64k on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.7462

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: 500

Training results

Training Loss	Epoch	Step	Validation Loss
36.613	0.0007	1	2.2585
18.6347	0.0353	50	1.1634
17.277	0.0707	100	0.9945
15.9303	0.1060	150	0.9258
12.4211	0.1413	200	0.8571
12.086	0.1767	250	0.8257
10.9912	0.2120	300	0.7918
11.3265	0.2474	350	0.7707
12.8793	0.2827	400	0.7552
11.8626	0.3180	450	0.7477
11.2528	0.3534	500	0.7462

Framework versions

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