# LLaMA [[llama]] ## 개요 [[overview]] LLaMA 모델은 Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin, Edouard Grave, Guillaume Lample에 의해 제안된 [LLaMA: Open and Efficient Foundation Language Models](https://huggingface.co/papers/2302.13971)에서 소개되었습니다. 이 모델은 7B에서 65B개의 파라미터까지 다양한 크기의 기초 언어 모델을 모아놓은 것입니다. 논문의 초록은 다음과 같습니다: *"LLaMA는 7B에서 65B개의 파라미터 수를 가진 기초 언어 모델의 모음입니다. 우리는 수조 개의 토큰으로 모델을 훈련시켰고, 공개적으로 이용 가능한 데이터셋만을 사용하여 최고 수준의 모델을 훈련시킬 수 있음을 보여줍니다. 특히, LLaMA-13B 모델은 대부분의 벤치마크에서 GPT-3 (175B)를 능가하며, LLaMA-65B는 최고 수준의 모델인 Chinchilla-70B와 PaLM-540B에 버금가는 성능을 보입니다. 우리는 모든 모델을 연구 커뮤니티에 공개합니다."* 팁: - LLaMA 모델의 가중치는 [이 양식](https://docs.google.com/forms/d/e/1FAIpQLSfqNECQnMkycAp2jP4Z9TFX0cGR4uf7b_fBxjY_OjhJILlKGA/viewform?usp=send_form)을 작성하여 얻을 수 있습니다. - 가중치를 다운로드한 후에는 이를 [변환 스크립트](https://github.com/huggingface/transformers/blob/main/src/transformers/models/llama/convert_llama_weights_to_hf.py)를 사용하여 Hugging Face Transformers 형식으로 변환해야합니다. 변환 스크립트를 실행하려면 아래의 예시 명령어를 참고하세요: ```bash 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 ``` - 변환을 하였다면 모델과 토크나이저는 다음과 같이 로드할 수 있습니다: ```python from transformers import LlamaForCausalLM, LlamaTokenizer tokenizer = LlamaTokenizer.from_pretrained("/output/path") model = LlamaForCausalLM.from_pretrained("/output/path") ``` 스크립트를 실행하기 위해서는 모델을 float16 정밀도로 전부 로드할 수 있을 만큼의 충분한 CPU RAM이 필요합니다. (가장 큰 버전의 모델이 여러 체크포인트로 나뉘어 있더라도, 각 체크포인트는 모델의 각 가중치의 일부를 포함하고 있기 때문에 모든 체크포인트를 RAM에 로드해야 합니다) 65B 모델의 경우, 총 130GB의 RAM이 필요합니다. - LLaMA 토크나이저는 [sentencepiece](https://github.com/google/sentencepiece)를 기반으로 하는 BPE 모델입니다. sentencepiece의 특징 중 하나는 시퀀스를 디코딩할 때 첫 토큰이 단어의 시작이라면 (예를 들어 "Banana"), 토크나이저는 문자열 앞에 공백을 추가하지 않는다는 것입니다. 이 모델은 [BlackSamorez](https://huggingface.co/BlackSamorez)의 기여와 함께, [zphang](https://huggingface.co/zphang)에 의해 제공되었습니다. Hugging Face에서의 구현 코드는 GPT-NeoX를 기반으로 하며 [여기](https://github.com/EleutherAI/gpt-neox)에서 찾을 수 있고, 저자의 코드 원본은 [여기](https://github.com/facebookresearch/llama)에서 확인할 수 있습니다. 원래 LLaMA 모델을 기반으로 Meta AI에서 몇 가지 후속 작업을 발표했습니다: - **Llama2**: Llama2는 구조적인 몇 가지 수정(Grouped Query Attention)을 통해 개선된 버전이며, 2조 개의 토큰으로 사전 훈련이 되어 있습니다. Llama2에 대한 자세한 내용은 [이 문서](llama2)를 참고하세요. ## 리소스 [[resources]] LLaMA를 시작하는 데 도움이 될 Hugging Face 및 커뮤니티(🌎로 표시)의 공식 자료 목록입니다. 여기에 자료를 제출하고 싶다면 Pull Request를 올려주세요! 추가할 자료는 기존의 자료와 중복되지 않고 새로운 내용을 보여주는 것이 좋습니다. - LLaMA 모델을 텍스트 분류 작업에 적용하기 위한 프롬프트 튜닝 방법에 대한 [노트북](https://colab.research.google.com/github/bigscience-workshop/petals/blob/main/examples/prompt-tuning-sst2.ipynb#scrollTo=f04ba4d2) 🌎 - [Stack Exchange](https://stackexchange.com/)에서 질문에 답하는 LLaMA를 훈련하는 방법을 위한 [StackLLaMA: RLHF로 LLaMA를 훈련하는 실전 가이드](https://huggingface.co/blog/stackllama#stackllama-a-hands-on-guide-to-train-llama-with-rlhf) 🌎 ⚗️ 최적화 - 제한된 메모리를 가진 GPU에서 xturing 라이브러리를 사용하여 LLaMA 모델을 미세 조정하는 방법에 대한 [노트북](https://colab.research.google.com/drive/1SQUXq1AMZPSLD4mk3A3swUIc6Y2dclme?usp=sharing) 🌎 ⚡️ 추론 - 🤗 PEFT 라이브러리의 PeftModel을 사용하여 LLaMA 모델을 실행하는 방법에 대한 [노트북](https://colab.research.google.com/github/DominguesM/alpaca-lora-ptbr-7b/blob/main/notebooks/02%20-%20Evaluate.ipynb) 🌎 - LangChain을 사용하여 PEFT 어댑터 LLaMA 모델을 로드하는 방법에 대한 [노트북](https://colab.research.google.com/drive/1l2GiSSPbajVyp2Nk3CFT4t3uH6-5TiBe?usp=sharing) 🌎 🚀 배포 - 🤗 PEFT 라이브러리와 사용자 친화적인 UI로 LLaMA 모델을 미세 조정하는 방법에 대한 [노트북](https://colab.research.google.com/github/lxe/simple-llama-finetuner/blob/master/Simple_LLaMA_FineTuner.ipynb#scrollTo=3PM_DilAZD8T) 🌎 - Amazon SageMaker에서 텍스트 생성을 위해 Open-LLaMA 모델을 배포하는 방법에 대한 [노트북](https://github.com/aws/amazon-sagemaker-examples/blob/main/introduction_to_amazon_algorithms/jumpstart-foundation-models/text-generation-open-llama.ipynb) 🌎 ## LlamaConfig [[llamaconfig]][[transformers.LlamaConfig]] #### transformers.LlamaConfig[[transformers.LlamaConfig]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/configuration_llama.py#L31) This is the configuration class to store the configuration of a LlamaModel. It is used to instantiate a Llama model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) Configuration objects inherit from [PreTrainedConfig](/docs/transformers/v5.8.0/ko/main_classes/configuration#transformers.PreTrainedConfig) and can be used to control the model outputs. Read the documentation from [PreTrainedConfig](/docs/transformers/v5.8.0/ko/main_classes/configuration#transformers.PreTrainedConfig) for more information. ```python >>> from transformers import LlamaModel, LlamaConfig >>> # Initializing a LLaMA llama-7b style configuration >>> configuration = LlamaConfig() >>> # Initializing a model from the llama-7b style configuration >>> model = LlamaModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` **Parameters:** vocab_size (`int`, *optional*, defaults to `32000`) : Vocabulary size of the model. Defines the number of different tokens that can be represented by the `input_ids`. hidden_size (`int`, *optional*, defaults to `4096`) : Dimension of the hidden representations. intermediate_size (`int`, *optional*, defaults to `11008`) : Dimension of the MLP representations. num_hidden_layers (`int`, *optional*, defaults to `32`) : Number of hidden layers in the Transformer decoder. num_attention_heads (`int`, *optional*, defaults to `32`) : Number of attention heads for each attention layer in the Transformer decoder. num_key_value_heads (`int`, *optional*) : This is the number of key_value heads that should be used to implement Grouped Query Attention. If `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed by meanpooling all the original heads within that group. For more details, check out [this paper](https://huggingface.co/papers/2305.13245). If it is not specified, will default to `num_attention_heads`. hidden_act (`str`, *optional*, defaults to `silu`) : The non-linear activation function (function or string) in the decoder. For example, `"gelu"`, `"relu"`, `"silu"`, etc. max_position_embeddings (`int`, *optional*, defaults to `2048`) : The maximum sequence length that this model might ever be used with. initializer_range (`float`, *optional*, defaults to `0.02`) : The standard deviation of the truncated_normal_initializer for initializing all weight matrices. rms_norm_eps (`float`, *optional*, defaults to `1e-06`) : The epsilon used by the rms normalization layers. use_cache (`bool`, *optional*, defaults to `True`) : Whether or not the model should return the last key/values attentions (not used by all models). Only relevant if `config.is_decoder=True` or when the model is a decoder-only generative model. pad_token_id (`int`, *optional*) : Token id used for padding in the vocabulary. bos_token_id (`int`, *optional*, defaults to `1`) : Token id used for beginning-of-stream in the vocabulary. eos_token_id (`Union[int, list[int]]`, *optional*, defaults to `2`) : Token id used for end-of-stream in the vocabulary. pretraining_tp (`int`, *optional*, defaults to `1`) : Experimental feature. Tensor parallelism rank used during pretraining. Please refer to [this document](https://huggingface.co/docs/transformers/main/perf_train_gpu_many#tensor-parallelism) to understand more about it. This value is necessary to ensure exact reproducibility of the pretraining results. Please refer to [this issue](https://github.com/pytorch/pytorch/issues/76232). tie_word_embeddings (`bool`, *optional*, defaults to `False`) : Whether to tie weight embeddings according to model's `tied_weights_keys` mapping. rope_parameters (`Union[~modeling_rope_utils.RopeParameters, dict]`, *optional*) : Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain a value for `rope_theta` and optionally parameters used for scaling in case you want to use RoPE with longer `max_position_embeddings`. attention_bias (`bool`, *optional*, defaults to `False`) : Whether to use a bias in the query, key, value and output projection layers during self-attention. attention_dropout (`Union[int, float]`, *optional*, defaults to `0.0`) : The dropout ratio for the attention probabilities. mlp_bias (`bool`, *optional*, defaults to `False`) : Whether to use a bias in up_proj, down_proj and gate_proj layers in the MLP layers. head_dim (`int`, *optional*) : The attention head dimension. If None, it will default to hidden_size // num_attention_heads ## LlamaTokenizer [[llamatokenizer]][[transformers.LlamaTokenizer]] #### transformers.LlamaTokenizer[[transformers.LlamaTokenizer]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/tokenization_llama.py#L39) Construct a Llama tokenizer. Based on byte-level Byte-Pair-Encoding. This uses notably ByteFallback and no normalization. ```python >>> from transformers import LlamaTokenizer >>> tokenizer = LlamaTokenizer.from_pretrained("hf-internal-testing/llama-tokenizer") >>> tokenizer.encode("Hello this is a test") [1, 15043, 445, 338, 263, 1243] ``` If you want to change the `bos_token` or the `eos_token`, make sure to specify them when initializing the model, or call `tokenizer.update_post_processor()` to make sure that the post-processing is correctly done (otherwise the values of the first token and final token of an encoded sequence will not be correct). For more details, checkout [post-processors] (https://huggingface.co/docs/tokenizers/api/post-processors) documentation. This tokenizer inherits from [PreTrainedTokenizerFast](/docs/transformers/v5.8.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) which contains most of the main methods. Users should refer to this superclass for more information regarding those methods. get_special_tokens_masktransformers.LlamaTokenizer.get_special_tokens_maskhttps://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/tokenization_utils_base.py#L1318[{"name": "token_ids_0", "val": ": list[int]"}, {"name": "token_ids_1", "val": ": list[int] | None = None"}, {"name": "already_has_special_tokens", "val": ": bool = False"}]- **token_ids_0** -- List of IDs for the (possibly already formatted) sequence. - **token_ids_1** -- Unused when `already_has_special_tokens=True`. Must be None in that case. - **already_has_special_tokens** -- Whether the sequence is already formatted with special tokens.0A list of integers in the range [0, 1]1 for a special token, 0 for a sequence token. Retrieve sequence ids from a token list that has no special tokens added. For fast tokenizers, data collators call this with `already_has_special_tokens=True` to build a mask over an already-formatted sequence. In that case, we compute the mask by checking membership in `all_special_ids`. **Parameters:** vocab (`str`, `dict` or `list`, *optional*) : Path to the vocabulary file, a dictionary or a list of tokens. merges (`str` or `list`, *optional*) : Path to the merges file or a list of merges. clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`) : Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts like extra spaces. unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this token instead. bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token. eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The end of sequence token. add_bos_token (`bool`, *optional*, defaults to `True`) : Whether or not to add an `bos_token` at the start of sequences. add_eos_token (`bool`, *optional*, defaults to `False`) : Whether or not to add an `eos_token` at the end of sequences. use_default_system_prompt (`bool`, *optional*, defaults to `False`) : Whether or not the default system prompt for Llama should be used add_prefix_space (`bool`, *optional*) : Whether or not the tokenizer should automatically add a prefix space **Returns:** `A list of integers in the range [0, 1]` 1 for a special token, 0 for a sequence token. #### save_vocabulary[[transformers.LlamaTokenizer.save_vocabulary]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/tokenization_utils_tokenizers.py#L509) ## LlamaTokenizerFast [[llamatokenizerfast]][[transformers.LlamaTokenizer]] #### transformers.LlamaTokenizer[[transformers.LlamaTokenizer]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/tokenization_llama.py#L39) Construct a Llama tokenizer. Based on byte-level Byte-Pair-Encoding. This uses notably ByteFallback and no normalization. ```python >>> from transformers import LlamaTokenizer >>> tokenizer = LlamaTokenizer.from_pretrained("hf-internal-testing/llama-tokenizer") >>> tokenizer.encode("Hello this is a test") [1, 15043, 445, 338, 263, 1243] ``` If you want to change the `bos_token` or the `eos_token`, make sure to specify them when initializing the model, or call `tokenizer.update_post_processor()` to make sure that the post-processing is correctly done (otherwise the values of the first token and final token of an encoded sequence will not be correct). For more details, checkout [post-processors] (https://huggingface.co/docs/tokenizers/api/post-processors) documentation. This tokenizer inherits from [PreTrainedTokenizerFast](/docs/transformers/v5.8.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) which contains most of the main methods. Users should refer to this superclass for more information regarding those methods. get_special_tokens_masktransformers.LlamaTokenizer.get_special_tokens_maskhttps://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/tokenization_utils_base.py#L1318[{"name": "token_ids_0", "val": ": list[int]"}, {"name": "token_ids_1", "val": ": list[int] | None = None"}, {"name": "already_has_special_tokens", "val": ": bool = False"}]- **token_ids_0** -- List of IDs for the (possibly already formatted) sequence. - **token_ids_1** -- Unused when `already_has_special_tokens=True`. Must be None in that case. - **already_has_special_tokens** -- Whether the sequence is already formatted with special tokens.0A list of integers in the range [0, 1]1 for a special token, 0 for a sequence token. Retrieve sequence ids from a token list that has no special tokens added. For fast tokenizers, data collators call this with `already_has_special_tokens=True` to build a mask over an already-formatted sequence. In that case, we compute the mask by checking membership in `all_special_ids`. **Parameters:** vocab (`str`, `dict` or `list`, *optional*) : Path to the vocabulary file, a dictionary or a list of tokens. merges (`str` or `list`, *optional*) : Path to the merges file or a list of merges. clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`) : Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts like extra spaces. unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this token instead. bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token. eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `""`) : The end of sequence token. add_bos_token (`bool`, *optional*, defaults to `True`) : Whether or not to add an `bos_token` at the start of sequences. add_eos_token (`bool`, *optional*, defaults to `False`) : Whether or not to add an `eos_token` at the end of sequences. use_default_system_prompt (`bool`, *optional*, defaults to `False`) : Whether or not the default system prompt for Llama should be used add_prefix_space (`bool`, *optional*) : Whether or not the tokenizer should automatically add a prefix space **Returns:** `A list of integers in the range [0, 1]` 1 for a special token, 0 for a sequence token. #### update_post_processor[[transformers.LlamaTokenizer.update_post_processor]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/tokenization_utils_tokenizers.py#L522) Updates the underlying post processor with the current `bos_token` and `eos_token`. #### save_vocabulary[[transformers.LlamaTokenizer.save_vocabulary]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/tokenization_utils_tokenizers.py#L509) ## LlamaModel [[llamamodel]][[transformers.LlamaModel]] #### transformers.LlamaModel[[transformers.LlamaModel]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/modeling_llama.py#L355) The bare Llama Model outputting raw hidden-states without any specific head on top. This model inherits from [PreTrainedModel](/docs/transformers/v5.8.0/ko/main_classes/model#transformers.PreTrainedModel). Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. forwardtransformers.LlamaModel.forwardhttps://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/modeling_llama.py#L372[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "past_key_values", "val": ": transformers.cache_utils.Cache | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "use_cache", "val": ": bool | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.8.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) - **attention_mask** (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) -- Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) - **position_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) - **past_key_values** (`~cache_utils.Cache`, *optional*) -- Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. - **inputs_embeds** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) -- Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. - **use_cache** (`bool`, *optional*) -- If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`).0[BaseModelOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPast) or `tuple(torch.FloatTensor)`A [BaseModelOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPast) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([LlamaConfig](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaConfig)) and inputs. The [LlamaModel](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaModel) forward method, overrides the `__call__` special method. Although the recipe for forward pass needs to be defined within this function, one should call the `Module` instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them. - **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model. If `past_key_values` is used only the last hidden-state of the sequences of shape `(batch_size, 1, hidden_size)` is output. - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if `config.is_encoder_decoder=True` in the cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. **Parameters:** config ([LlamaConfig](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaConfig)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.8.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights. **Returns:** `[BaseModelOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPast) or `tuple(torch.FloatTensor)`` A [BaseModelOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPast) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([LlamaConfig](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaConfig)) and inputs. ## LlamaForCausalLM [[llamaforcausallm]][[transformers.LlamaForCausalLM]] #### transformers.LlamaForCausalLM[[transformers.LlamaForCausalLM]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/modeling_llama.py#L429) The Llama Model for causal language modeling. This model inherits from [PreTrainedModel](/docs/transformers/v5.8.0/ko/main_classes/model#transformers.PreTrainedModel). Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. forwardtransformers.LlamaForCausalLM.forwardhttps://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/modeling_llama.py#L443[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "past_key_values", "val": ": transformers.cache_utils.Cache | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "use_cache", "val": ": bool | None = None"}, {"name": "logits_to_keep", "val": ": int | torch.Tensor = 0"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.8.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) - **attention_mask** (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) -- Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) - **position_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) - **past_key_values** (`~cache_utils.Cache`, *optional*) -- Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. - **inputs_embeds** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) -- Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. - **labels** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`. - **use_cache** (`bool`, *optional*) -- If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`). - **logits_to_keep** (`Union[int, torch.Tensor]`, *optional*, defaults to `0`) -- If an `int`, compute logits for the last `logits_to_keep` tokens. If `0`, calculate logits for all `input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that token can save memory, which becomes pretty significant for long sequences or large vocabulary size. If a `torch.Tensor`, must be 1D corresponding to the indices to keep in the sequence length dimension. This is useful when using packed tensor format (single dimension for batch and sequence length).0[CausalLMOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.CausalLMOutputWithPast) or `tuple(torch.FloatTensor)`A [CausalLMOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.CausalLMOutputWithPast) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([LlamaConfig](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaConfig)) and inputs. The [LlamaForCausalLM](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) forward method, overrides the `__call__` special method. Although the recipe for forward pass needs to be defined within this function, one should call the `Module` instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them. - **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Language modeling loss (for next-token prediction). - **logits** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) -- Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax). - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Example: ```python >>> from transformers import AutoTokenizer, LlamaForCausalLM >>> model = LlamaForCausalLM.from_pretrained("meta-llama/Llama-2-7b-hf") >>> tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-2-7b-hf") >>> prompt = "Hey, are you conscious? Can you talk to me?" >>> inputs = tokenizer(prompt, return_tensors="pt") >>> # Generate >>> generate_ids = model.generate(inputs.input_ids, max_length=30) >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0] "Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you." ``` **Parameters:** config ([LlamaForCausalLM](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaForCausalLM)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.8.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights. **Returns:** `[CausalLMOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.CausalLMOutputWithPast) or `tuple(torch.FloatTensor)`` A [CausalLMOutputWithPast](/docs/transformers/v5.8.0/ko/main_classes/output#transformers.modeling_outputs.CausalLMOutputWithPast) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([LlamaConfig](/docs/transformers/v5.8.0/ko/model_doc/llama2#transformers.LlamaConfig)) and inputs. ## LlamaForSequenceClassification [[llamaforsequenceclassification]][[transformers.LlamaForSequenceClassification]] #### transformers.LlamaForSequenceClassification[[transformers.LlamaForSequenceClassification]] [Source](https://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/models/llama/modeling_llama.py#L502) forwardtransformers.LlamaForSequenceClassification.forwardhttps://github.com/huggingface/transformers/blob/v5.8.0/src/transformers/modeling_layers.py#L110[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "past_key_values", "val": ": transformers.cache_utils.Cache | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "use_cache", "val": ": bool | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.8.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) - **attention_mask** (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) -- Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) - **position_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) - **past_key_values** (`~cache_utils.Cache`, *optional*) -- Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. - **inputs_embeds** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) -- Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. - **labels** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`. - **use_cache** (`bool`, *optional*) -- If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`).0`SequenceClassifierOutputWithPast` or `tuple(torch.FloatTensor)`A `SequenceClassifierOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration (`None`) and inputs. The `GenericForSequenceClassification` forward method, overrides the `__call__` special method. Although the recipe for forward pass needs to be defined within this function, one should call the `Module` instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them. - **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Classification (or regression if config.num_labels==1) loss. - **logits** (`torch.FloatTensor` of shape `(batch_size, config.num_labels)`) -- Classification (or regression if config.num_labels==1) scores (before SoftMax). - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. **Parameters:** input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) : Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.8.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.8.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) : Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) : Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) past_key_values (`~cache_utils.Cache`, *optional*) : Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.8.0/ko/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) : Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) : Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`. use_cache (`bool`, *optional*) : If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`). **Returns:** ``SequenceClassifierOutputWithPast` or `tuple(torch.FloatTensor)`` A `SequenceClassifierOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration (`None`) and inputs.