Startseite Erkunden Hilfe
Registrieren Anmelden
lainlives
/
sft-trainer
1
0
Fork 0
Dateien Issues 0 Pull-Requests 0 Wiki

Keine Beschreibung

Struktur: 7995893a21
Branches Tags
sft-trainer
ZIP TAR.GZ
llc 7995893a21 Update config.yaml vor 3 Monaten
scripts 93538163f5 Deploy Qwen3-Coder-Next uncensored fine-tuner vor 4 Monaten
.gitattributes 1cf7fc58ca initial commit vor 4 Monaten
.gitignore 712316c058 Deploy Qwen3-Coder-Next uncensored fine-tuner vor 4 Monaten
Dockerfile d820e41318 Dockerfile: install unsloth, remove old deps vor 4 Monaten
README.md 712316c058 Deploy Qwen3-Coder-Next uncensored fine-tuner vor 4 Monaten
app.py d9138d2db8 Update app.py vor 3 Monaten
config.yaml 7995893a21 Update config.yaml vor 3 Monaten
requirements.txt f35dacf5b8 Slim requirements (unsloth bundles deps) vor 4 Monaten
train.py bc8de7d068 Update train.py to Unsloth FastModel vor 4 Monaten

README.md

title: Qwen3-Coder-Next Uncensored Fine-Tuning emoji: "\U0001F525" colorFrom: red colorTo: yellow sdk: docker app_port: 7860 suggested_hardware: a100-large suggested_storage: large pinned: true

license: apache-2.0

Qwen3-Coder-Next Uncensored Fine-Tuner

Fine-tune Qwen/Qwen3-Coder-Next (80B MoE / 3B active) using QLoRA to remove refusal behavior.

Setup

  1. Create a HF Space with Docker SDK and a100-large hardware
  2. Add Secrets in Space Settings:
    • HF_TOKEN — your Hugging Face write token
    • WANDB_API_KEY — (optional) for training metrics logging
  3. Upload this repo to the Space
  4. The Gradio UI will let you configure and launch training

Features

  • QLoRA 4-bit fine-tuning (fits on single A100 80GB)
  • Multiple uncensored dataset options built-in
  • Custom dataset upload support
  • Real-time training metrics in UI
  • One-click LoRA merge + push to Hub
  • Abliteration support (no training needed)

Hardware Requirements

Method VRAM Recommended HF Hardware
QLoRA 4-bit ~45-60GB A100 Large (80GB)
Abliteration ~45-60GB A100 Large (80GB)