# Fast-Inference with Ctranslate2

Speedup inference while reducing memory by 2x-4x using int8 inference in C++ on CPU or GPU.

pip install hf-hub-ctranslate2>=2.0.8 ctranslate2>=3.14.0

Converted on 2023-06-07 using

ct2-transformers-converter --model tiiuae/falcon-7b --output_dir /home/michael/tmp-ct2fast-falcon-7b --force --copy_files tokenizer.json README.md tokenizer_config.json generation_config.json special_tokens_map.json .gitattributes --quantization int8_float16 --trust_remote_code

Checkpoint compatible to ctranslate2>=3.15.0 and hf-hub-ctranslate2>=2.0.8

compute_type=int8_float16 for device="cuda"
compute_type=int8 for device="cpu"

from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer

model_name = "michaelfeil/ct2fast-falcon-7b"
# use either TranslatorCT2fromHfHub or GeneratorCT2fromHfHub here, depending on model.
model = GeneratorCT2fromHfHub(
        # load in int8 on CUDA
        model_name_or_path=model_name,
        device="cuda",
        compute_type="int8_float16",
        # tokenizer=AutoTokenizer.from_pretrained("tiiuae/falcon-7b")
)
outputs = model.generate(
    text=["def fibonnaci(", "User: How are you doing? Bot:"],
    max_length=64,
    include_prompt_in_result=False
)
print(outputs)

Licence and other remarks:

This is just a quantized version. Licence conditions are intended to be idential to original huggingface repo.

Original description

🚀 Falcon-7B

Falcon-7B is a 7B parameters causal decoder-only model built by TII and trained on 1,500B tokens of RefinedWeb enhanced with curated corpora. It is made available under the Apache 2.0 license.

Paper coming soon 😊.

Why use Falcon-7B?

It outperforms comparable open-source models (e.g., MPT-7B, StableLM, RedPajama etc.), thanks to being trained on 1,500B tokens of RefinedWeb enhanced with curated corpora. See the OpenLLM Leaderboard.
It features an architecture optimized for inference, with FlashAttention (Dao et al., 2022) and multiquery (Shazeer et al., 2019).
It is made available under a permissive Apache 2.0 license allowing for commercial use, without any royalties or restrictions.

⚠️ This is a raw, pretrained model, which should be further finetuned for most usecases. If you are looking for a version better suited to taking generic instructions in a chat format, we recommend taking a look at Falcon-7B-Instruct.

🔥 Looking for an even more powerful model? Falcon-40B is Falcon-7B's big brother!

from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

model = "tiiuae/falcon-7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
    device_map="auto",
)
sequences = pipeline(
   "Girafatron is obsessed with giraffes, the most glorious animal on the face of this Earth. Giraftron believes all other animals are irrelevant when compared to the glorious majesty of the giraffe.\nDaniel: Hello, Girafatron!\nGirafatron:",
    max_length=200,
    do_sample=True,
    top_k=10,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
)
for seq in sequences:
    print(f"Result: {seq['generated_text']}")

💥 Falcon LLMs require PyTorch 2.0 for use with transformers!

Model Card for Falcon-7B

Model Details

Model Description

Developed by: https://www.tii.ae;
Model type: Causal decoder-only;
Language(s) (NLP): English and French;
License: Apache 2.0.

Model Source

Paper: coming soon.

Uses

Direct Use

Research on large language models; as a foundation for further specialization and finetuning for specific usecases (e.g., summarization, text generation, chatbot, etc.)

Out-of-Scope Use

Production use without adequate assessment of risks and mitigation; any use cases which may be considered irresponsible or harmful.

Bias, Risks, and Limitations

Falcon-7B is trained on English and French data only, and will not generalize appropriately to other languages. Furthermore, as it is trained on a large-scale corpora representative of the web, it will carry the stereotypes and biases commonly encountered online.

Recommendations

We recommend users of Falcon-7B to consider finetuning it for the specific set of tasks of interest, and for guardrails and appropriate precautions to be taken for any production use.

How to Get Started with the Model

from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

model = "tiiuae/falcon-7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
    device_map="auto",
)
sequences = pipeline(
   "Girafatron is obsessed with giraffes, the most glorious animal on the face of this Earth. Giraftron believes all other animals are irrelevant when compared to the glorious majesty of the giraffe.\nDaniel: Hello, Girafatron!\nGirafatron:",
    max_length=200,
    do_sample=True,
    top_k=10,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
)
for seq in sequences:
    print(f"Result: {seq['generated_text']}")

Training Details

Training Data

Falcon-7B was trained on 1,500B tokens of RefinedWeb, a high-quality filtered and deduplicated web dataset which we enhanced with curated corpora. Significant components from our curated copora were inspired by The Pile (Gao et al., 2020).

Data source	Fraction	Tokens	Sources
RefinedWeb-English	79%	1,185B	massive web crawl
Books	7%	110B
Conversations	6%	85B	Reddit, StackOverflow, HackerNews
Code	3%	45B
RefinedWeb-French	3%	45B	massive web crawl
Technical	2%	30B	arXiv, PubMed, UPSTO, etc.

The data was tokenized with the Falcon-7B/40B tokenizer.

Training Procedure

Falcon-7B was trained on 384 A100 40GB GPUs, using a 2D parallelism strategy (PP=2, DP=192) combined with ZeRO.

Training Hyperparameters

Hyperparameter	Value	Comment
Precision	`bfloat16`
Optimizer	AdamW
Learning rate	6e-4	4B tokens warm-up, cosine decay to 1.2e-5
Weight decay	1e-1
Z-loss	1e-4
Batch size	2304	30B tokens ramp-up

Speeds, Sizes, Times

Training happened in early March 2023 and took about two weeks.

Evaluation

Paper coming soon.

See the OpenLLM Leaderboard for early results.

Technical Specifications

Model Architecture and Objective

Falcon-7B is a causal decoder-only model trained on a causal language modeling task (i.e., predict the next token).

The architecture is broadly adapted from the GPT-3 paper (Brown et al., 2020), with the following differences:

Positionnal embeddings: rotary (Su et al., 2021);
Attention: multiquery (Shazeer et al., 2019) and FlashAttention (Dao et al., 2022);
Decoder-block: parallel attention/MLP with a single layer norm.

Hyperparameter	Value	Comment
Layers	32
`d_model`	4544	Increased to compensate for multiquery
`head_dim`	64	Reduced to optimise for FlashAttention
Vocabulary	65024
Sequence length	2048

Compute Infrastructure

Hardware

Falcon-7B was trained on AWS SageMaker, on 384 A100 40GB GPUs in P4d instances.

Software

Falcon-7B was trained a custom distributed training codebase, Gigatron. It uses a 3D parallelism approach combined with ZeRO and high-performance Triton kernels (FlashAttention, etc.)

Citation

Paper coming soon 😊. In the meanwhile, you can use the following information to cite:

@article{falcon40b,
  title={{Falcon-40B}: an open large language model with state-of-the-art performance},
  author={Almazrouei, Ebtesam and Alobeidli, Hamza and Alshamsi, Abdulaziz and Cappelli, Alessandro and Cojocaru, Ruxandra and Debbah, Merouane and Goffinet, Etienne and Heslow, Daniel and Launay, Julien and Malartic, Quentin and Noune, Badreddine and Pannier, Baptiste and Penedo, Guilherme},
  year={2023}
}

To learn more about the pretraining dataset, see the 📓 RefinedWeb paper.

@article{refinedweb,
  title={The {R}efined{W}eb dataset for {F}alcon {LLM}: outperforming curated corpora with web data, and web data only},
  author={Guilherme Penedo and Quentin Malartic and Daniel Hesslow and Ruxandra Cojocaru and Alessandro Cappelli and Hamza Alobeidli and Baptiste Pannier and Ebtesam Almazrouei and Julien Launay},
  journal={arXiv preprint arXiv:2306.01116},
  eprint={2306.01116},
  eprinttype = {arXiv},
  url={https://arxiv.org/abs/2306.01116},
  year={2023}
}

License

Falcon-7B is made available under the Apache 2.0 license.

Contact

falconllm@tii.ae

michaelfeil
/

ct2fast-falcon-7b