transformers/docs/source/en/model_doc/code_llama.md

CodeLlama

Overview

The CodeLlama model was proposed in Code Llama: Open Foundation Models for Code by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.

The abstract from the paper is the following:

We release Code Llama, a family of large language models for code based on Llama 2 providing state-of-the-art performance among open models, infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks. We provide multiple flavors to cover a wide range of applications: foundation models (Code Llama), Python specializations (Code Llama - Python), and instruction-following models (Code Llama - Instruct) with 7B, 13B and 34B parameters each. All models are trained on sequences of 16k tokens and show improvements on inputs with up to 100k tokens. 7B and 13B Code Llama and Code Llama - Instruct variants support infilling based on surrounding content. Code Llama reaches state-of-the-art performance among open models on several code benchmarks, with scores of up to 53% and 55% on HumanEval and MBPP, respectively. Notably, Code Llama - Python 7B outperforms Llama 2 70B on HumanEval and MBPP, and all our models outperform every other publicly available model on MultiPL-E. We release Code Llama under a permissive license that allows for both research and commercial use.

Checkout all CodeLlama models here

The Llama2 family models, on which Code Llama is based, were trained using bfloat16, but the original inference uses float16. The checkpoints uploaded on the hub use torch_dtype = 'float16'which will be used by theAutoModelAPI to cast the checkpoints fromtorch.float32totorch.float16`.

The dtype of the online weights is mostly irrelevant, unless you are using torch_dtype="auto" when initializing a model using model = AutoModelForCausalLM.from_pretrained("path", torch_dtype = "auto"). The reason is that the model will first be downloaded ( using the dtype of the checkpoints online) then it will be casted to the default dtype of torch (becomes torch.float32) and finally, if there is a torch_dtype provided in the config, it will be used.

Training the model in float16 is not recommended and known to produce nan, as such the model should be trained in bfloat16.

Tips:

• These models have the same architcture as the Llama2 models
• The infilling task is supported out of the box. You should be using the tokenizer.fill_token where you want your input to be filled.
• The model conversion script is the same as for the Llama2 familly:

Here is a sample usage

python src/transformers/models/llama/convert_llama_weights_to_hf.py \
    --input_dir /path/to/downloaded/llama/weights --model_size 7B --output_dir /output/path

• After conversion, the model and tokenizer can be loaded via:

>>> from transformers import LlamaForCausalLM, CodeLlamaTokenizer

>>> tokenizer = CodeLlamaTokenizer.from_pretrained("codellama/CodeLlama-7b-hf")
>>> model = LlamaForCausalLM.from_pretrained("codellama/CodeLlama-7b-hf")
>>> PROMPT = '''def remove_non_ascii(s: str) -> str:
    """ <FILL_ME>
    return result
'''
>>> input_ids = tokenizer(PROMPT, return_tensors="pt")["input_ids"]
>>> generated_ids = model.generate(input_ids, max_new_tokens=128)

>>> filling = tokenizer.batch_decode(generated_ids[:, input_ids.shape[1]:], skip_special_tokens = True)[0]
>>> print(PROMPT.replace("<FILL_ME>", filling))
def remove_non_ascii(s: str) -> str:
    """ Remove non-ASCII characters from a string.

    Args:
        s: The string to remove non-ASCII characters from.

    Returns:
        The string with non-ASCII characters removed.
    """
    result = ""
    for c in s:
        if ord(c) < 128:
            result += c
    return result

If you only want the infilled part:

>>> from transformers import pipeline
>>> import torch

>>> generator = pipeline("text-generation",model="codellama/CodeLlama-7b-hf",torch_dtype=torch.float16, device_map="auto")
>>> generator('def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return result', max_new_tokens = 128, return_type = 1)

Note that executing the script requires enough CPU RAM to host the whole model in float16 precision (even if the biggest versions come in several checkpoints they each contain a part of each weight of the model, so we need to load them all in RAM). For the 75B model, it's thus 145GB of RAM needed.

• The LLaMA tokenizer is a BPE model based on sentencepiece. One quirk of sentencepiece is that when decoding a sequence, if the first token is the start of the word (e.g. "Banana"), the tokenizer does not prepend the prefix space to the string.

This model was contributed by ArthurZucker. The original code of the authors can be found here.

CodeLlamaTokenizer

autodoc CodeLlamaTokenizer - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences - save_vocabulary

CodeLlamaTokenizerFast

autodoc CodeLlamaTokenizerFast - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences - update_post_processor - save_vocabulary