Instructions to use debisoft/mpt-7b-8k-instruct-peft-compatible with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use debisoft/mpt-7b-8k-instruct-peft-compatible with Transformers:

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

pipe = pipeline("text-generation", model="debisoft/mpt-7b-8k-instruct-peft-compatible", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("debisoft/mpt-7b-8k-instruct-peft-compatible", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("debisoft/mpt-7b-8k-instruct-peft-compatible", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use debisoft/mpt-7b-8k-instruct-peft-compatible with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "debisoft/mpt-7b-8k-instruct-peft-compatible"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "debisoft/mpt-7b-8k-instruct-peft-compatible",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/debisoft/mpt-7b-8k-instruct-peft-compatible

SGLang

How to use debisoft/mpt-7b-8k-instruct-peft-compatible 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 "debisoft/mpt-7b-8k-instruct-peft-compatible" \
    --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": "debisoft/mpt-7b-8k-instruct-peft-compatible",
		"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 "debisoft/mpt-7b-8k-instruct-peft-compatible" \
        --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": "debisoft/mpt-7b-8k-instruct-peft-compatible",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use debisoft/mpt-7b-8k-instruct-peft-compatible with Docker Model Runner:
```
docker model run hf.co/debisoft/mpt-7b-8k-instruct-peft-compatible
```

debisoft commited on Dec 9, 2023

Commit

a43d8a8

1 Parent(s): defbb61

Added output_attentions: bool=False to GroupedQueryAttention.forward() as a temporary fix for AWQ

Browse files

Files changed (1) hide show

attention.py +1 -1

attention.py CHANGED Viewed

@@ -260,7 +260,7 @@ class GroupedQueryAttention(nn.Module):
         self.out_proj = FC_CLASS_REGISTRY[fc_type](self.d_model, self.d_model, **fc_kwargs)
         self.out_proj._is_residual = True
-    def forward(self, x: torch.Tensor, past_key_value: Optional[Tuple[torch.Tensor, torch.Tensor]]=None, attn_bias: Optional[torch.Tensor]=None, attention_mask: Optional[torch.Tensor]=None, is_causal: bool=True, needs_weights: bool=False) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor, torch.Tensor]]]:
         qkv = self.Wqkv(x)
         if self.clip_qkv:
             qkv = qkv.clamp(min=-self.clip_qkv, max=self.clip_qkv)

         self.out_proj = FC_CLASS_REGISTRY[fc_type](self.d_model, self.d_model, **fc_kwargs)
         self.out_proj._is_residual = True
+    def forward(self, x: torch.Tensor, past_key_value: Optional[Tuple[torch.Tensor, torch.Tensor]]=None, attn_bias: Optional[torch.Tensor]=None, attention_mask: Optional[torch.Tensor]=None, is_causal: bool=True, output_attentions: bool=False, needs_weights: bool=False) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor, torch.Tensor]]]:
         qkv = self.Wqkv(x)
         if self.clip_qkv:
             qkv = qkv.clamp(min=-self.clip_qkv, max=self.clip_qkv)