Transformers documentation

Mistral

Transformers

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v4.40.1).

Join the Hugging Face community

and get access to the augmented documentation experience

Collaborate on models, datasets and Spaces

Faster examples with accelerated inference

Switch between documentation themes

to get started

Mistral

Overview

Mistral was introduced in the this blogpost by Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.

The introduction of the blog post says:

Mistral AI team is proud to release Mistral 7B, the most powerful language model for its size to date.

Mistral-7B is the first large language model (LLM) released by mistral.ai.

Architectural details

Mistral-7B is a decoder-only Transformer with the following architectural choices:

Sliding Window Attention - Trained with 8k context length and fixed cache size, with a theoretical attention span of 128K tokens
GQA (Grouped Query Attention) - allowing faster inference and lower cache size.
Byte-fallback BPE tokenizer - ensures that characters are never mapped to out of vocabulary tokens.

For more details refer to the release blog post.

License

Mistral-7B is released under the Apache 2.0 license.

Usage tips

The Mistral team has released 3 checkpoints:

a base model, Mistral-7B-v0.1, which has been pre-trained to predict the next token on internet-scale data.
an instruction tuned model, Mistral-7B-Instruct-v0.1, which is the base model optimized for chat purposes using supervised fine-tuning (SFT) and direct preference optimization (DPO).
an improved instruction tuned model, Mistral-7B-Instruct-v0.2, which improves upon v1.

The base model can be used as follows:

>>> from transformers import AutoModelForCausalLM, AutoTokenizer

>>> model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1", device_map="auto")
>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")

>>> prompt = "My favourite condiment is"

>>> model_inputs = tokenizer([prompt], return_tensors="pt").to("cuda")
>>> model.to(device)

>>> generated_ids = model.generate(**model_inputs, max_new_tokens=100, do_sample=True)
>>> tokenizer.batch_decode(generated_ids)[0]
"My favourite condiment is to ..."

The instruction tuned model can be used as follows:

>>> from transformers import AutoModelForCausalLM, AutoTokenizer

>>> model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2", device_map="auto")
>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")

>>> messages = [
...     {"role": "user", "content": "What is your favourite condiment?"},
...     {"role": "assistant", "content": "Well, I'm quite partial to a good squeeze of fresh lemon juice. It adds just the right amount of zesty flavour to whatever I'm cooking up in the kitchen!"},
...     {"role": "user", "content": "Do you have mayonnaise recipes?"}
... ]

>>> model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to("cuda")

>>> generated_ids = model.generate(model_inputs, max_new_tokens=100, do_sample=True)
>>> tokenizer.batch_decode(generated_ids)[0]
"Mayonnaise can be made as follows: (...)"

As can be seen, the instruction-tuned model requires a chat template to be applied to make sure the inputs are prepared in the right format.

Speeding up Mistral by using Flash Attention

The code snippets above showcase inference without any optimization tricks. However, one can drastically speed up the model by leveraging Flash Attention, which is a faster implementation of the attention mechanism used inside the model.

First, make sure to install the latest version of Flash Attention 2 to include the sliding window attention feature.

pip install -U flash-attn --no-build-isolation

Make also sure that you have a hardware that is compatible with Flash-Attention 2. Read more about it in the official documentation of the flash attention repository. Make also sure to load your model in half-precision (e.g. torch.float16)

To load and run a model using Flash Attention-2, refer to the snippet below:

>>> import torch
>>> from transformers import AutoModelForCausalLM, AutoTokenizer

>>> model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1", torch_dtype=torch.float16, attn_implementation="flash_attention_2", device_map="auto")
>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")

>>> prompt = "My favourite condiment is"

>>> model_inputs = tokenizer([prompt], return_tensors="pt").to("cuda")
>>> model.to(device)

>>> generated_ids = model.generate(**model_inputs, max_new_tokens=100, do_sample=True)
>>> tokenizer.batch_decode(generated_ids)[0]
"My favourite condiment is to (...)"

Expected speedups

Below is a expected speedup diagram that compares pure inference time between the native implementation in transformers using mistralai/Mistral-7B-v0.1 checkpoint and the Flash Attention 2 version of the model.

Sliding window Attention

The current implementation supports the sliding window attention mechanism and memory efficient cache management. To enable sliding window attention, just make sure to have a flash-attn version that is compatible with sliding window attention (>=2.3.0).

The Flash Attention-2 model uses also a more memory efficient cache slicing mechanism - as recommended per the official implementation of Mistral model that use rolling cache mechanism we keep the cache size fixed (self.config.sliding_window), support batched generation only for padding_side="left" and use the absolute position of the current token to compute the positional embedding.

Shrinking down Mistral using quantization

As the Mistral model has 7 billion parameters, that would require about 14GB of GPU RAM in half precision (float16), since each parameter is stored in 2 bytes. However, one can shrink down the size of the model using quantization. If the model is quantized to 4 bits (or half a byte per parameter),that requires only about 3.5GB of RAM.

Quantizing a model is as simple as passing a quantization_config to the model. Below, we’ll leverage the BitsAndyBytes quantization (but refer to this page for other quantization methods):

>>> import torch
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig

>>> # specify how to quantize the model
>>> quantization_config = BitsAndBytesConfig(
...         load_in_4bit=True,
...         bnb_4bit_quant_type="nf4",
...         bnb_4bit_compute_dtype="torch.float16",
... )

>>> model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2", quantization_config=True, device_map="auto")
>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")

>>> prompt = "My favourite condiment is"

>>> messages = [
...     {"role": "user", "content": "What is your favourite condiment?"},
...     {"role": "assistant", "content": "Well, I'm quite partial to a good squeeze of fresh lemon juice. It adds just the right amount of zesty flavour to whatever I'm cooking up in the kitchen!"},
...     {"role": "user", "content": "Do you have mayonnaise recipes?"}
... ]

>>> model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to("cuda")

>>> generated_ids = model.generate(model_inputs, max_new_tokens=100, do_sample=True)
>>> tokenizer.batch_decode(generated_ids)[0]
"The expected output"

This model was contributed by Younes Belkada and Arthur Zucker . The original code can be found here.

Resources

A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with Mistral. If you’re interested in submitting a resource to be included here, please feel free to open a Pull Request and we’ll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource.

Text Generation

A demo notebook to perform supervised fine-tuning (SFT) of Mistral-7B can be found here. 🌎
A blog post on how to fine-tune LLMs in 2024 using Hugging Face tooling. 🌎
The Alignment Handbook by Hugging Face includes scripts and recipes to perform supervised fine-tuning (SFT) and direct preference optimization with Mistral-7B. This includes scripts for full fine-tuning, QLoRa on a single GPU as well as multi-GPU fine-tuning.
Causal language modeling task guide

MistralConfig

class transformers.MistralConfig

< source >

( vocab_size = 32000 hidden_size = 4096 intermediate_size = 14336 num_hidden_layers = 32 num_attention_heads = 32 num_key_value_heads = 8 hidden_act = 'silu' max_position_embeddings = 131072 initializer_range = 0.02 rms_norm_eps = 1e-06 use_cache = True pad_token_id = None bos_token_id = 1 eos_token_id = 2 tie_word_embeddings = False rope_theta = 10000.0 sliding_window = 4096 attention_dropout = 0.0 **kwargs )

Parameters

vocab_size (int, optional, defaults to 32000) — Vocabulary size of the Mistral model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling MistralModel
hidden_size (int, optional, defaults to 4096) — Dimension of the hidden representations.
intermediate_size (int, optional, defaults to 14336) — Dimension of the MLP representations.
num_hidden_layers (int, optional, defaults to 32) — Number of hidden layers in the Transformer encoder.
num_attention_heads (int, optional, defaults to 32) — Number of attention heads for each attention layer in the Transformer encoder.
num_key_value_heads (int, optional, defaults to 8) — 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 checkout [this paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to 8`.
hidden_act (str or function, optional, defaults to "silu") — The non-linear activation function (function or string) in the decoder.
max_position_embeddings (int, optional, defaults to 4096*32) — The maximum sequence length that this model might ever be used with. Mistral’s sliding window attention allows sequence of up to 4096*32 tokens.
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.
pad_token_id (int, optional) — The id of the padding token.
bos_token_id (int, optional, defaults to 1) — The id of the “beginning-of-sequence” token.
eos_token_id (int, optional, defaults to 2) — The id of the “end-of-sequence” token.
tie_word_embeddings (bool, optional, defaults to False) — Whether the model’s input and output word embeddings should be tied.
rope_theta (float, optional, defaults to 10000.0) — The base period of the RoPE embeddings.
sliding_window (int, optional, defaults to 4096) — Sliding window attention window size. If not specified, will default to 4096.
attention_dropout (float, optional, defaults to 0.0) — The dropout ratio for the attention probabilities.

This is the configuration class to store the configuration of a MistralModel. It is used to instantiate an Mistral 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 Mistral-7B-v0.1 or Mistral-7B-Instruct-v0.1.

mistralai/Mistral-7B-v0.1 mistralai/Mistral-7B-Instruct-v0.1

Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.

>>> from transformers import MistralModel, MistralConfig

>>> # Initializing a Mistral 7B style configuration
>>> configuration = MistralConfig()

>>> # Initializing a model from the Mistral 7B style configuration
>>> model = MistralModel(configuration)

>>> # Accessing the model configuration
>>> configuration = model.config

MistralModel

class transformers.MistralModel

< source >

( config: MistralConfig )

Parameters

config (MistralConfig) — 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() method to load the model weights. config — MistralConfig

The bare Mistral Model outputting raw hidden-states without any specific head on top. This model inherits from 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Transformer decoder consisting of config.num_hidden_layers layers. Each layer is a MistralDecoderLayer

forward

< source >

( input_ids: LongTensor = None attention_mask: Optional = None position_ids: Optional = None past_key_values: Optional = None inputs_embeds: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None )

Parameters

input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide it.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are 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?

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

If you want to change padding behavior, you should read modeling_opt._prepare_decoder_attention_mask and modify to your needs. See diagram 1 in the paper for more information on the default strategy.
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
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?
past_key_values (Cache or tuple(tuple(torch.FloatTensor)), 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.

Two formats are allowed:
- a Cache instance;
- Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)). This is also known as the legacy cache format.
The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.

If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) 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).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

The MistralModel 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.

MistralForCausalLM

class transformers.MistralForCausalLM

< source >

( config )

forward

< source >

( input_ids: LongTensor = None attention_mask: Optional = None position_ids: Optional = None past_key_values: Optional = None inputs_embeds: Optional = None labels: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.CausalLMOutputWithPast or tuple(torch.FloatTensor)

Parameters

input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide it.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are 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?

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

If you want to change padding behavior, you should read modeling_opt._prepare_decoder_attention_mask and modify to your needs. See diagram 1 in the paper for more information on the default strategy.
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
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?
past_key_values (Cache or tuple(tuple(torch.FloatTensor)), 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.

Two formats are allowed:
- a Cache instance;
- Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)). This is also known as the legacy cache format.
The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.

If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) 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).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Args — 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].

Returns

transformers.modeling_outputs.CausalLMOutputWithPast or tuple(torch.FloatTensor)

A 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 (MistralConfig) and inputs.

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 (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head))

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.

The MistralForCausalLM forward method, overrides the __call__ special method.

Example:

>>> from transformers import AutoTokenizer, MistralForCausalLM

>>> model = MistralForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1")
>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")

>>> 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."

MistralForSequenceClassification

class transformers.MistralForSequenceClassification

< source >

( config )

Parameters

config (MistralConfig) — 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() method to load the model weights.

The Mistral Model transformer with a sequence classification head on top (linear layer).

MistralForSequenceClassification uses the last token in order to do the classification, as other causal models (e.g. GPT-2) do.

Since it does classification on the last token, it requires to know the position of the last token. If a pad_token_id is defined in the configuration, it finds the last token that is not a padding token in each row. If no pad_token_id is defined, it simply takes the last value in each row of the batch. Since it cannot guess the padding tokens when inputs_embeds are passed instead of input_ids, it does the same (take the last value in each row of the batch).

This model inherits from 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< source >

Parameters

input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide it.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are 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?

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

If you want to change padding behavior, you should read modeling_opt._prepare_decoder_attention_mask and modify to your needs. See diagram 1 in the paper for more information on the default strategy.
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
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?
past_key_values (Cache or tuple(tuple(torch.FloatTensor)), 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.

Two formats are allowed:
- a Cache instance;
- Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)). This is also known as the legacy cache format.
The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.

If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) 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).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.
labels (torch.LongTensor of shape (batch_size,), optional) — Labels for computing the sequence classification/regression loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels == 1 a regression loss is computed (Mean-Square loss), If config.num_labels > 1 a classification loss is computed (Cross-Entropy).

The MistralForSequenceClassification forward method, overrides the __call__ special method.

FlaxMistralModel

class transformers.FlaxMistralModel

< source >

( config: MistralConfig input_shape: Tuple = (1, 1) seed: int = 0 dtype: dtype = <class 'jax.numpy.float32'> _do_init: bool = True **kwargs )

Parameters

config (MistralConfig) — 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() method to load the model weights.
dtype (jax.numpy.dtype, optional, defaults to jax.numpy.float32) — The data type of the computation. Can be one of jax.numpy.float32, jax.numpy.float16, or jax.numpy.bfloat16.

This can be used to enable mixed-precision training or half-precision inference on GPUs or TPUs. If specified all the computation will be performed with the given dtype.

Note that this only specifies the dtype of the computation and does not influence the dtype of model parameters.

If you wish to change the dtype of the model parameters, see to_fp16() and to_bf16().

The bare Mistral Model transformer outputting raw hidden-states without any specific head on top.

This model inherits from FlaxPreTrainedModel. 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 Flax Linen flax.nn.Module subclass. Use it as a regular Flax Module and refer to the Flax documentation for all matter related to general usage and behavior.

Finally, this model supports inherent JAX features such as:

call

< source >

( input_ids attention_mask = None position_ids = None params: dict = None past_key_values: dict = None dropout_rng: PRNGKey = None train: bool = False output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_flax_outputs.FlaxBaseModelOutputWithPast or tuple(torch.FloatTensor)

Parameters

input_ids (numpy.ndarray of shape (batch_size, input_ids_length)) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide it.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are input IDs?
attention_mask (numpy.ndarray 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?

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

If you want to change padding behavior, you should read modeling_opt._prepare_decoder_attention_mask and modify to your needs. See diagram 1 in the paper for more information on the default strategy.
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
position_ids (numpy.ndarray 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?
past_key_values (Dict[str, np.ndarray], optional, returned by init_cache or when passing previous past_key_values) — Dictionary of pre-computed hidden-states (key and values in the attention blocks) that can be used for fast auto-regressive decoding. Pre-computed key and value hidden-states are of shape [batch_size, max_length].
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.modeling_flax_outputs.FlaxBaseModelOutputWithPast or tuple(torch.FloatTensor)

A transformers.modeling_flax_outputs.FlaxBaseModelOutputWithPast 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 (MistralConfig) and inputs.

last_hidden_state (jnp.ndarray of shape (batch_size, sequence_length, hidden_size)) — Sequence of hidden-states at the output of the last layer of the model.
past_key_values (Dict[str, jnp.ndarray]) — Dictionary of pre-computed hidden-states (key and values in the attention blocks) that can be used for fast auto-regressive decoding. Pre-computed key and value hidden-states are of shape [batch_size, max_length].
hidden_states (tuple(jnp.ndarray), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of jnp.ndarray (one for the output of the embeddings + 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 initial embedding outputs.
attentions (tuple(jnp.ndarray), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of jnp.ndarray (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.

The FlaxMistralPreTrainedModel forward method, overrides the __call__ special method.

This example uses a random model as the real ones are all very big. To get proper results, you should use mistralai/Mistral-7B-v0.1 instead of ksmcg/Mistral-tiny. If you get out-of-memory when loading that checkpoint, you can try adding device_map="auto" in the from_pretrained call.

Example:

>>> from transformers import AutoTokenizer, FlaxMistralModel

>>> tokenizer = AutoTokenizer.from_pretrained("ksmcg/Mistral-tiny")
>>> model = FlaxMistralModel.from_pretrained("ksmcg/Mistral-tiny")

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="jax")
>>> outputs = model(**inputs)

>>> last_hidden_states = outputs.last_hidden_state

FlaxMistralForCausalLM

class transformers.FlaxMistralForCausalLM

< source >

( config: MistralConfig input_shape: Tuple = (1, 1) seed: int = 0 dtype: dtype = <class 'jax.numpy.float32'> _do_init: bool = True **kwargs )

Parameters

config (MistralConfig) — 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() method to load the model weights.
dtype (jax.numpy.dtype, optional, defaults to jax.numpy.float32) — The data type of the computation. Can be one of jax.numpy.float32, jax.numpy.float16, or jax.numpy.bfloat16.

This can be used to enable mixed-precision training or half-precision inference on GPUs or TPUs. If specified all the computation will be performed with the given dtype.

Note that this only specifies the dtype of the computation and does not influence the dtype of model parameters.

If you wish to change the dtype of the model parameters, see to_fp16() and to_bf16().

The Mistral Model transformer with a language modeling head (linear layer) on top.

This model is also a Flax Linen flax.nn.Module subclass. Use it as a regular Flax Module and refer to the Flax documentation for all matter related to general usage and behavior.

Finally, this model supports inherent JAX features such as:

call

< source >

( input_ids attention_mask = None position_ids = None params: dict = None past_key_values: dict = None dropout_rng: PRNGKey = None train: bool = False output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_flax_outputs.FlaxCausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)

Parameters

input_ids (numpy.ndarray of shape (batch_size, input_ids_length)) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide it.

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

What are input IDs?
attention_mask (numpy.ndarray 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?

Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

If you want to change padding behavior, you should read modeling_opt._prepare_decoder_attention_mask and modify to your needs. See diagram 1 in the paper for more information on the default strategy.
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
position_ids (numpy.ndarray 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?
past_key_values (Dict[str, np.ndarray], optional, returned by init_cache or when passing previous past_key_values) — Dictionary of pre-computed hidden-states (key and values in the attention blocks) that can be used for fast auto-regressive decoding. Pre-computed key and value hidden-states are of shape [batch_size, max_length].
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.modeling_flax_outputs.FlaxCausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)

A transformers.modeling_flax_outputs.FlaxCausalLMOutputWithCrossAttentions 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 (MistralConfig) and inputs.

logits (jnp.ndarray of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
hidden_states (tuple(jnp.ndarray), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of jnp.ndarray (one for the output of the embeddings + 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 initial embedding outputs.
attentions (tuple(jnp.ndarray), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of jnp.ndarray (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.
cross_attentions (tuple(jnp.ndarray), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of jnp.ndarray (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Cross attentions weights after the attention softmax, used to compute the weighted average in the cross-attention heads.
past_key_values (tuple(tuple(jnp.ndarray)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of jnp.ndarray tuples of length config.n_layers, with each tuple containing the cached key, value states of the self-attention and the cross-attention layers if model is used in encoder-decoder setting. Only relevant if config.is_decoder = True.

Contains pre-computed hidden-states (key and values in the attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

The FlaxMistralPreTrainedModel forward method, overrides the __call__ special method.

Example:

>>> from transformers import AutoTokenizer, FlaxMistralForCausalLM

>>> tokenizer = AutoTokenizer.from_pretrained("ksmcg/Mistral-tiny")
>>> model = FlaxMistralForCausalLM.from_pretrained("ksmcg/Mistral-tiny")

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="np")
>>> outputs = model(**inputs)

>>> # retrieve logts for next token
>>> next_token_logits = outputs.logits[:, -1]

< > Update on GitHub

←MegatronGPT2 Mixtral→