metadata

tags:
  - pytorch
  - safetensors
  - transformers
  - gpt-oss
  - multilingual
  - text-generation
language:
  - en
  - es
  - fr
  - de
  - it
  - pt
license: apache-2.0
model_type: gpt-oss
pipeline_tag: text-generation
base_model: openai/gpt-oss-20b

GPT-OSS-20B Function Calling Model

This repository contains the GPT-OSS-20B model fine-tuned on function calling data in PyTorch/Safetensors format, ready for use with the Transformers library.

Model Details

Base Model: openai/gpt-oss-20b
Fine-tuning Dataset: Salesforce/xlam-function-calling-60k (2000 samples)
Fine-tuning Method: LoRA (r=8, alpha=16)
Context Length: 131,072 tokens
Model Size: 20B parameters

Files

model.safetensors: Model weights in Safetensors format
config.json: Model configuration
tokenizer.json, tokenizer_config.json: Tokenizer files
generation_config.json: Generation configuration

Usage

With Transformers Library

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained(
    "cuijian0819/gpt-oss-20b-function-calling",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("cuijian0819/gpt-oss-20b-function-calling")

# Generate text
inputs = tokenizer("Your prompt here", return_tensors="pt")
outputs = model.generate(**inputs, max_length=100, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Fine-tuning

This model can be further fine-tuned using standard PyTorch/Transformers workflows:

from transformers import AutoModelForCausalLM, AutoTokenizer, Trainer, TrainingArguments

model = AutoModelForCausalLM.from_pretrained("cuijian0819/gpt-oss-20b-function-calling")
tokenizer = AutoTokenizer.from_pretrained("cuijian0819/gpt-oss-20b-function-calling")

# Your fine-tuning code here

GGUF Version

For efficient inference with llama.cpp or Ollama, check out the GGUF version: cuijian0819/gpt-oss-20b-function-calling-gguf

Training Details

Training Epochs: 5
Learning Rate: 0.0002
Batch Size: 4
Gradient Accumulation: 4
Max Length: 1024

License

This model inherits the license from the base openai/gpt-oss-20b model.

Citation

@misc{gpt-oss-20b-function-calling,
  title={GPT-OSS-20B Function Calling Model},
  author={cuijian0819},
  year={2025},
  url={https://huggingface.co/cuijian0819/gpt-oss-20b-function-calling}
}