Instructions to use Changahou/Llama8B_mathinstruct_SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Changahou/Llama8B_mathinstruct_SFT with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("/home/chenjh2/disk_chenjh/models/models/Llama-3-8b-Instruct") model = PeftModel.from_pretrained(base_model, "Changahou/Llama8B_mathinstruct_SFT") - Notebooks
- Google Colab
- Kaggle
File size: 841 Bytes
05dbd61 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | adapter_name_or_path: /home/chenjh2/disk_chenjh/models
bf16: true
cutoff_len: 2048
dataset: OpenR1-Math-220k
dataset_dir: data
ddp_timeout: 180000000
do_train: true
enable_thinking: true
finetuning_type: lora
flash_attn: auto
gradient_accumulation_steps: 8
include_num_input_tokens_seen: true
learning_rate: 5.0e-05
logging_steps: 5
lora_alpha: 16
lora_dropout: 0
lora_rank: 8
lora_target: all
lr_scheduler_type: cosine
max_grad_norm: 1.0
max_samples: 100000
model_name_or_path: /home/chenjh2/disk_chenjh/models/models/Llama-3-8b-Instruct
num_train_epochs: 3.0
optim: adamw_torch
output_dir: saves/Llama-3-8B-Instruct/lora/train_2025-07-09-13-09-41
packing: false
per_device_train_batch_size: 2
plot_loss: true
preprocessing_num_workers: 16
report_to: none
save_steps: 100
stage: sft
template: llama3
trust_remote_code: true
warmup_steps: 0
|