Hugging Face's logo

mgfrantz
/
axolotl-test

PEFT
PyTorch
Safetensors
GGUF
llama
Generated from Trainer
Inference Endpoints
4-bit precision
bitsandbytes
conversational
Model card Files Community
axolotl-test / README.md
mgfrantz's picture
mgfrantz
End of training
6a931ec verified
preview code
|
raw
history blame
3.79 kB
metadata
base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
library_name: peft
license: apache-2.0
tags:
  - axolotl
  - generated_from_trainer
model-index:
  - name: axolotl-test
    results: []

Built with Axolotl

See axolotl config

axolotl version: 0.4.1


# Model config
adapter: qlora
base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
# base_model: meta-llama/Llama-3.2-3B
bf16: auto

# HF hub config (push to huggingface)
# requires HF_TOKEN api key to be set (👈🔑secrets)
hf_use_auth_token: true
hub_model_id: mgfrantz/axolotl-test
mlflow_experiment_name: axolotl-test

# # Data config
dataset_prepared_path: data
chat_template: chatml
datasets:
  - path: data/train.jsonl
    ds_type: json
    data_files:
      - data/train.jsonl
    conversation: alpaca
    type: sharegpt

test_datasets:
  - path: data/eval.jsonl
    ds_type: json
    # You need to specify a split. For "json" datasets the default split is called "train".
    split: train
    type: sharegpt
    conversation: alpaca
    data_files:
      - data/eval.jsonl


# Training config
debug: null
deepspeed: null
early_stopping_patience: null
eval_sample_packing: false
evals_per_epoch: 4
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false


learning_rate: 0.0002
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lora_target_modules: null
lr_scheduler: cosine
micro_batch_size: 8
model_type: LlamaForCausalLM
num_epochs: 4
optimizer: paged_adamw_32bit
output_dir: ./outputs/qlora-out
pad_to_sequence_len: true
resume_from_checkpoint: null
sample_packing: true
saves_per_epoch: 1
sequence_len: 4096
special_tokens: null
strict: false
tf32: false
tokenizer_type: LlamaTokenizer
train_on_inputs: false
wandb_entity: null
wandb_log_model: null
wandb_name: null
wandb_project: null
wandb_watch: null
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

axolotl-test

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.4338

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: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 10
  • num_epochs: 4

Training results

Training Loss Epoch Step Validation Loss
3.4962 0.5714 1 2.4779
5.3564 1.0714 2 2.4760
4.3272 1.6429 3 2.4633
4.7348 2.1429 4 2.4338

Framework versions

  • PEFT 0.13.2
  • Transformers 4.45.2
  • Pytorch 2.4.0+cu121
  • Datasets 3.0.1
  • Tokenizers 0.20.1