Instructions to use adamo1139/Yi-34B-200K-HESOYAM-2206 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use adamo1139/Yi-34B-200K-HESOYAM-2206 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="adamo1139/Yi-34B-200K-HESOYAM-2206")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("adamo1139/Yi-34B-200K-HESOYAM-2206")
model = AutoModelForCausalLM.from_pretrained("adamo1139/Yi-34B-200K-HESOYAM-2206")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use adamo1139/Yi-34B-200K-HESOYAM-2206 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "adamo1139/Yi-34B-200K-HESOYAM-2206"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "adamo1139/Yi-34B-200K-HESOYAM-2206",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/adamo1139/Yi-34B-200K-HESOYAM-2206

SGLang

How to use adamo1139/Yi-34B-200K-HESOYAM-2206 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 "adamo1139/Yi-34B-200K-HESOYAM-2206" \
    --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": "adamo1139/Yi-34B-200K-HESOYAM-2206",
		"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 "adamo1139/Yi-34B-200K-HESOYAM-2206" \
        --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": "adamo1139/Yi-34B-200K-HESOYAM-2206",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use adamo1139/Yi-34B-200K-HESOYAM-2206 with Docker Model Runner:
```
docker model run hf.co/adamo1139/Yi-34B-200K-HESOYAM-2206
```

Basic Model Info

1 epoch on adamo1139/uninstruct-v1-experimental-chatml and then 1 epoch on adamo1139/HESOYAM_v0.3. I used GaLore for both stages.

This is a model trained on only human data, finetuned to behave like a person on 4chan board /x/ or redditor. Data used has comments from 1 4chan board "paranormal" and about 10 reddit subreddits. There's also a pippa in case you want to roleplay. Have a look at dataset to know what to expect.

Use ChatML prompt format with a system prompt like those in adamo1139/HESOYAM_v0.3, so A chat on 4chan or A chat on subreddit /r/wallstreetbets. It behaves like OpenAI slopped model with system prompt A chat so I advise you to avoid using that.

Downloads last month: 6

Safetensors

Model size

34B params

Tensor type

F16

Datasets used to train adamo1139/Yi-34B-200K-HESOYAM-2206

Paper for adamo1139/Yi-34B-200K-HESOYAM-2206

GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection

Paper • 2403.03507 • Published Mar 6, 2024 • 190