zhanxing

afm-qwen3-8b-sft-fiannce-music-movie

Deploy Dedicated

README

License: apache-2.0

Training

Base model: Qwen/Qwen3-8B
Dataset: 26,384 AFM trace SFT examples
Method: 4-bit QLoRA followed by LoRA merge into the full base model
Context length: 4096
Effective batch size: 16
Learning rate: 2e-4 with linear decay
Epochs: 2
Framework: LLaMA-Factory

Final training loss: 0.1007. This is a training-only metric, not an independent evaluation score.

Usage

Load zhanxing/afm-qwen3-8b-sft-fiannce-music-movie directly with AutoTokenizer.from_pretrained and AutoModelForCausalLM.from_pretrained. Use the Qwen3 chat template and provide required tool definitions in the prompt.

Limitations

Evaluate on held-out tool-use tasks before production use. The model may reproduce errors or biases present in the tool-trace training data.

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Container

Run this model inference with full control and performance in your environment.

Learn more

Model Details

Model Provider

zhanxing

Model Tree

Base

Qwen/Qwen3-8B

Fine-tuned

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated EndpointsContainer

Explore FriendliAI today

Get started Talk to an engineer