Text Generation
Transformers
PyTorch
Safetensors
English
Korean
gpt2
generation
text-generation-inference
Instructions to use psyche/kogpt with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use psyche/kogpt with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="psyche/kogpt")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("psyche/kogpt") model = AutoModelForCausalLM.from_pretrained("psyche/kogpt", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use psyche/kogpt with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "psyche/kogpt" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "psyche/kogpt", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/psyche/kogpt
- SGLang
How to use psyche/kogpt with 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 "psyche/kogpt" \ --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": "psyche/kogpt", "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 "psyche/kogpt" \ --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": "psyche/kogpt", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use psyche/kogpt with Docker Model Runner:
docker model run hf.co/psyche/kogpt
Update README.md
Browse files
README.md
CHANGED
|
@@ -5,48 +5,6 @@ language:
|
|
| 5 |
tags:
|
| 6 |
- generation
|
| 7 |
license: apache-2.0
|
| 8 |
-
model-index:
|
| 9 |
-
- name: task_1
|
| 10 |
-
results:
|
| 11 |
-
- task:
|
| 12 |
-
type: natural-language-generation
|
| 13 |
-
dataset:
|
| 14 |
-
type: hellaswag
|
| 15 |
-
name: hellaswag(10 shots)
|
| 16 |
-
metrics:
|
| 17 |
-
- type: acc_norm
|
| 18 |
-
value: 27.7
|
| 19 |
-
- name: task_2
|
| 20 |
-
results:
|
| 21 |
-
- task:
|
| 22 |
-
type: natural-language-generation
|
| 23 |
-
dataset:
|
| 24 |
-
type: ARC
|
| 25 |
-
name: ARC(25 shots)
|
| 26 |
-
metrics:
|
| 27 |
-
- type: acc_norm
|
| 28 |
-
value: 23.8
|
| 29 |
-
- name: task_3
|
| 30 |
-
results:
|
| 31 |
-
- task:
|
| 32 |
-
type: natural-language-generation
|
| 33 |
-
dataset:
|
| 34 |
-
type: MMLU
|
| 35 |
-
name: MMLU(5 shots)
|
| 36 |
-
metrics:
|
| 37 |
-
- type: acc
|
| 38 |
-
value: 24.9
|
| 39 |
-
|
| 40 |
-
- name: task_4
|
| 41 |
-
results:
|
| 42 |
-
- task:
|
| 43 |
-
type: natural-language-generation
|
| 44 |
-
dataset:
|
| 45 |
-
type: TruthfulQA
|
| 46 |
-
name: TruthfulQA(0 shots)
|
| 47 |
-
metrics:
|
| 48 |
-
- type: mc2
|
| 49 |
-
value: 46.5
|
| 50 |
---
|
| 51 |
|
| 52 |
Pretrained GPT2 with expanded n_ctx up to 2048(also with expanded embedding dimension to 1536) in Korean.
|
|
|
|
| 5 |
tags:
|
| 6 |
- generation
|
| 7 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
Pretrained GPT2 with expanded n_ctx up to 2048(also with expanded embedding dimension to 1536) in Korean.
|