adapter: qlora
auto_resume_from_checkpoints: true
base_model: openlm-research/open_llama_3b
bf16: true
chat_template: llama3
dataset_prepared_path: null
dataset_processes: 1
datasets:
- data_files:
  - 5668b44a803035dd_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/5668b44a803035dd_train_data.json
  type:
    field_input: concepts
    field_instruction: instruction
    field_output: response
    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: 200
eval_table_size: null
evals_per_epoch:
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: false
group_by_length: false
hub_model_id: error577/b04780ef-9f9b-425c-8e3a-41c0a82a1156
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
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:
micro_batch_size: 1
mlflow_experiment_name: /tmp/5668b44a803035dd_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 4
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: 200
sequence_len: 512
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: eeda891a-7fde-4fb3-9e00-351e4d608c46
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: eeda891a-7fde-4fb3-9e00-351e4d608c46
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

b04780ef-9f9b-425c-8e3a-41c0a82a1156

This model is a fine-tuned version of openlm-research/open_llama_3b on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4543

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: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
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: 30
num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss
0.8702	0.0001	1	0.9949
0.6079	0.0160	200	0.4871
0.4528	0.0319	400	0.4786
0.2824	0.0479	600	0.4708
0.3962	0.0639	800	0.4652
0.476	0.0798	1000	0.4598
0.5171	0.0958	1200	0.4604
0.4526	0.1118	1400	0.4589
0.3575	0.1277	1600	0.4584
0.4238	0.1437	1800	0.4533
0.3328	0.1597	2000	0.4513
0.3756	0.1756	2200	0.4549
0.4138	0.1916	2400	0.4504
0.3911	0.2075	2600	0.4506
0.4822	0.2235	2800	0.4514
0.4627	0.2395	3000	0.4543

Framework versions

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