How to use from
SGLang
Install from pip and serve model
# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "12sciencejnv/TalkGPT" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "12sciencejnv/TalkGPT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'
Use Docker images
docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "12sciencejnv/TalkGPT" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "12sciencejnv/TalkGPT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'
Quick Links

TalkGPT

This model is a fine-tuned version of BlenderBot-400M (distilled) based on a custom conversational dataset. It is designed to generate conversational responses in English.

License

Apache 2.0

Datasets

The model is fine-tuned on a custom dataset consisting of conversational dialogues.

Language

English

Metrics

  • BLEU: 0.1687 (calculated on the validation set)
  • ROUGE-1: 0.4078
  • ROUGE-2: 0.1912
  • ROUGE-L: 0.3418
  • ROUGE-Lsum: 0.3401
  • Training Loss: 0.2460 (final training loss after fine-tuning)

Base Model

The model is based on the BlenderBot-400M-distill architecture by Facebook AI.

Pipeline Tag

text-generation

Library Name

transformers

Tags

BlenderBot, Conversational, Fine-tuned, Text Generation

Eval Results

The model achieved the following results on the validation set:

  • BLEU: 0.1687
  • ROUGE-1: 0.4078
  • ROUGE-2: 0.1912
  • ROUGE-L: 0.3418
  • ROUGE-Lsum: 0.3401
  • Training Loss: 0.2460 after 3 epochs of fine-tuning.
Downloads last month
19
Safetensors
Model size
0.4B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for 12sciencejnv/TalkGPT

Finetuned
(21)
this model

Evaluation results