adapter: lora
base_model: samoline/b81a070e-182b-4067-b1f5-9353758a2de2
bf16: true
chat_template: llama3
dataloader_num_workers: 2
dataloader_pin_memory: true
dataset_prepared_path: null
datasets:
- data_files:
  - 122edb1ca65d6b0e_dataset.json
  ds_type: json
  format: custom
  path: /workspace/input_data/
  type:
    field_instruction: instruction
    field_output: output
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
ddp_timeout: 7200
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 2
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 1
gradient_checkpointing: true
group_by_length: true
hub_model_id: rene-contango/15006d89-44b8-4531-b9a2-9b9f4ff70c8a
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 10
lora_alpha: 64
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: null
micro_batch_size: 8
mlflow_experiment_name: /tmp/122edb1ca65d6b0e_dataset.json
model_type: AutoModelForCausalLM
num_epochs: 1
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
save_safetensors: true
saves_per_epoch: 2
sequence_len: 2048
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_disabled: true
warmup_steps: 50
weight_decay: 0.01
xformers_attention: null

15006d89-44b8-4531-b9a2-9b9f4ff70c8a

This model is a fine-tuned version of samoline/b81a070e-182b-4067-b1f5-9353758a2de2 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6348
Memory/max Mem Active(gib): 49.23
Memory/max Mem Allocated(gib): 49.23
Memory/device Mem Reserved(gib): 56.87

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: 8
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 64
total_eval_batch_size: 64
optimizer: Use OptimizerNames.ADAMW_TORCH 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: 50
training_steps: 2209

Training results

Training Loss	Epoch	Step	Validation Loss	Mem Active(gib)	Mem Allocated(gib)	Mem Reserved(gib)
No log	0	0	0.6438	39.41	39.41	47.89
0.6447	0.5002	1105	0.6348	49.23	49.23	56.87

Framework versions

PEFT 0.17.0
Transformers 4.55.0
Pytorch 2.6.0+cu118
Datasets 4.0.0
Tokenizers 0.21.4