Hugging Face's logo

Pclanglais
/
Jambert

Text Generation
Transformers
TensorBoard
Safetensors
French
jamba
axolotl
custom_code
Inference Endpoints
4-bit precision
bitsandbytes
Model card Files Metrics Community
1
Jambert / adapter /README.md
Pclanglais's picture
Pclanglais
Upload folder using huggingface_hub
c215d89 verified
preview code
|
raw
history blame contribute delete
No virus
2.76 kB
metadata
library_name: peft
tags:
  - generated_from_trainer
base_model: jamba
model-index:
  - name: out
    results: []

Built with Axolotl

See axolotl config

axolotl version: 0.4.0

base_model: jamba
trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: scikit_admin_result.json
    ds_type: json
    type: sharegpt
    conversation: chatml
dataset_prepared_path:
val_set_size: 0.01
output_dir: ./out

sequence_len: 6000
sample_packing: true
pad_to_sequence_len: false
eval_sample_packing: true

use_wandb: false

adapter: qlora
lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true

low_cpu_mem_usage: true
gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 2
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch: 2
saves_per_epoch: 2
debug:
weight_decay: 0.0
special_tokens:

out

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2356

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 10
  • num_epochs: 2

Training results

Training Loss Epoch Step Validation Loss
0.4337 0.0 1 0.3783
0.2537 0.5 103 0.2345
0.2161 1.0 206 0.2258
0.1821 1.47 309 0.2356

Framework versions

  • PEFT 0.10.0
  • Transformers 4.40.0.dev0
  • Pytorch 2.2.2+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.0