adapter: lora
base_model: facebook/opt-350m
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 1141106889a52def_train_data.json
  ds_type: json
  field: prompt
  path: /workspace/input_data/1141106889a52def_train_data.json
  type: completion
debug: null
deepspeed: null
device_map: auto
early_stopping_patience: 3
ema_decay: 0.9992
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
greater_is_better: false
group_by_length: false
hub_model_id: JoshMe1/2022fb42-410a-423f-80fd-80b188c091c5
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: true
local_rank: null
logging_steps: 10
lora_alpha: 256
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: reduce_lr_on_plateau
lr_scheduler_factor: 0.5
lr_scheduler_patience: 2
max_memory:
  0: 130GB
max_steps: 384
metric_for_best_model: eval_loss
micro_batch_size: 4
mlflow_experiment_name: /tmp/1141106889a52def_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 1
optimizer: adamw_hf
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
saves_per_epoch: null
sequence_len: 2048
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
use_ema: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: fd67c684-5868-4130-9282-57785255ae6d
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: fd67c684-5868-4130-9282-57785255ae6d
warmup_steps: 200
weight_decay: 0.01
xformers_attention: null

2022fb42-410a-423f-80fd-80b188c091c5

This model is a fine-tuned version of facebook/opt-350m on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.2299

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: 0.0001
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_HF with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: reduce_lr_on_plateau
lr_scheduler_warmup_steps: 200
training_steps: 384

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0005	1	2.2588
5.7038	0.0548	100	1.2997
5.081	0.1095	200	1.2565
4.8973	0.1643	300	1.2299

Framework versions

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