tiiuae
/

falcon-40b-instruct

Text Generation

text-generation-inference

Model card Files Files and versions

Instructions to use tiiuae/falcon-40b-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiiuae/falcon-40b-instruct with Transformers:

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

pipe = pipeline("text-generation", model="tiiuae/falcon-40b-instruct", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-40b-instruct", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("tiiuae/falcon-40b-instruct", trust_remote_code=True)

Notebooks
Google Colab
Kaggle
Local Apps

How to use tiiuae/falcon-40b-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiiuae/falcon-40b-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiiuae/falcon-40b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tiiuae/falcon-40b-instruct

How to use tiiuae/falcon-40b-instruct 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 "tiiuae/falcon-40b-instruct" \
    --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": "tiiuae/falcon-40b-instruct",
		"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 "tiiuae/falcon-40b-instruct" \
        --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": "tiiuae/falcon-40b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tiiuae/falcon-40b-instruct with Docker Model Runner:
```
docker model run hf.co/tiiuae/falcon-40b-instruct
```

Resources

View closed (29)

Custom 4-bit Finetuning 5-7 times faster inference than QLora

#9 opened almost 3 years ago by

Install & run tiiuae/falcon-40b-instruct easily using llmpm

#95 opened 3 months ago by

Update README.md

#94 opened 4 months ago by

I'm not fine with quality of this model compared to the size in billions of parameters.

#93 opened about 1 year ago by

Update tokenizer_config.json

#92 opened over 1 year ago by

Update tokenizer_config.json

#91 opened over 1 year ago by

falcon-40b-instruct error on Inference endpoint while deploying

#90 opened over 1 year ago by

AI World

#89 opened about 2 years ago by

MohammadMuzamil

Adding `safetensors` variant of this model

#88 opened over 2 years ago by

combining falcon 40b instruct with langchain

#87 opened over 2 years ago by

Update generation_config.json

#85 opened over 2 years ago by

Update generation_config.json

#84 opened over 2 years ago by

Getting gibberish output with Falcon-40b instruct

#83 opened over 2 years ago by

Falcon 40B Inference on GKE Autopilot A100 40GB

#82 opened over 2 years ago by

Adding `safetensors` variant of this model

#81 opened over 2 years ago by

Adding `safetensors` variant of this model

#80 opened over 2 years ago by

CPU or GPU

#76 opened almost 3 years ago by

Optimizing Inference Time for Chat Conversations on Falcon

#73 opened almost 3 years ago by

Use input attention mask instead of casual mask in attention

#72 opened almost 3 years ago by

is there a way to not use trust_remote = True

#71 opened almost 3 years ago by

Unable to load and run finetuned falcon model

#70 opened almost 3 years ago by

Parameters contains nan numbers when loading model locally

#69 opened almost 3 years ago by

ValueError: sharded is not supported for AutoModel ERROR

#68 opened almost 3 years ago by

ValueError in KoboldAI when loading the model

#66 opened almost 3 years ago by

Cannot set "instructions" when invoking inference endpoint

#65 opened almost 3 years ago by

Changes in modelling_RW.py to be able to handle past_key_values for faster model generations

#64 opened almost 3 years ago by

Model sometimes generates '</s>'

#63 opened almost 3 years ago by

Correct blogpost link

#62 opened almost 3 years ago by

Error: ShardCannotStart

#61 opened almost 3 years ago by

Finetuning Falcon-40B-Instruct For ChatBot Use Case

#59 opened almost 3 years ago by

Adding `safetensors` variant of this model

#58 opened almost 3 years ago by

Add `tokenizer_class` to get `pipeline` to load tokenizer

#57 opened almost 3 years ago by

Adding `safetensors` variant of this model

#56 opened almost 3 years ago by

ValueError: Error raised by inference API: Model tiiuae/falcon-40b-instruct time out using HuggingFaceHub

#55 opened almost 3 years ago by

Question about Apache 2.0 license

#54 opened almost 3 years ago by

Running the Falcon-40B-Instruct model on Azure Kubernetes Service

#53 opened almost 3 years ago by

Experimental ggml demos

#52 opened almost 3 years ago by

Truncated output from API call through langchain

#51 opened almost 3 years ago by deleted

Experiences with complex instructions

#50 opened almost 3 years ago by

Update README.md

#49 opened almost 3 years ago by

Why Rotary Positional Embeddings Over Alibi?

#48 opened almost 3 years ago by

About Input validation error: `inputs` tokens + `max_new_tokens` must be <= 1512.

#47 opened almost 3 years ago by

is Alibi version available for fine tuning to a large context window?

#46 opened almost 3 years ago by

Finetune Falcon-4b with large token size.

#44 opened almost 3 years ago by

Model returns entire input prompt together with output

#43 opened almost 3 years ago by

Instruction prompt

#42 opened almost 3 years ago by

Update README.md

#41 opened almost 3 years ago by

Arabic Language support

#40 opened almost 3 years ago by

Request: DOI

#39 opened almost 3 years ago by

what is the input token length of Falcon-40B and -7B models?

#38 opened almost 3 years ago by