Mistral
Last updated
Last updated
Mistral-7B-v0.1 is Mistral AI’s first Large Language Model (LLM).
Mistral-7B-v0.1 is a decoder-based LM 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.
We also provide an instruction fine-tuned model: Mistral-7B-Instruct-v0.1 which can be used for chat-based inference.
For more details please read our
Both Mistral-7B-v0.1 and Mistral-7B-Instruct-v0.1 are released under the Apache 2.0 license.
Mistral-7B-v0.1 and Mistral-7B-Instruct-v0.1 can be found on the
These ready-to-use checkpoints can be downloaded and used via the BOINCAI Hub:
Copied
Raw weights for Mistral-7B-v0.1 and Mistral-7B-Instruct-v0.1 can be downloaded from:
Mistral-7B-v0.1
Mistral-7B-Instruct-v0.1
To use these raw checkpoints with HuggingFace you can use the convert_mistral_weights_to_hf.py script to convert them to the HuggingFace format:
Copied
You can then load the converted model from the output/path:
Copied
First, make sure to install the latest version of Flash Attention 2 to include the sliding window attention feature.
Copied
To load and run a model using Flash Attention 2, refer to the snippet below:
Copied
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.
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.
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.
( vocab_size = 32000hidden_size = 4096intermediate_size = 14336num_hidden_layers = 32num_attention_heads = 32num_key_value_heads = 8hidden_act = 'silu'max_position_embeddings = 131072initializer_range = 0.02rms_norm_eps = 1e-06use_cache = Truepad_token_id = Nonebos_token_id = 1eos_token_id = 2tie_word_embeddings = Falserope_theta = 10000.0sliding_window = 4096**kwargs )
Parameters
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 409632) — The maximum sequence length that this model might ever be used with. Mistral’s sliding window attention allows sequence of up to 409632 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-12) — 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.
tie_word_embeddings(bool, optional, defaults to False) — Whether to tie weight embeddings
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.
Example —
Copied
( config: MistralConfig )
Parameters
Transformer decoder consisting of config.num_hidden_layers layers. Each layer is a MistralDecoderLayer
forward
( input_ids: LongTensor = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: typing.Optional[typing.List[torch.FloatTensor]] = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = 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.
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.
If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).
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].
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)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_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.
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.
( config )
forward
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.
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.
If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).
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].
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)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_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.
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
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.
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.
Example:
Copied
( config )
Parameters
The Mistral Model transformer with a sequence classification head on top (linear layer).
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).
forward
( input_ids: LongTensor = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: typing.Optional[typing.List[torch.FloatTensor]] = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Nonelabels: typing.Optional[torch.LongTensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = 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.
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.
If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).
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].
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)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_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.
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).
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.
Make also sure that you have a hardware that is compatible with Flash-Attention 2. Read more about it in the official documentation of repository. Make also sure to load your model in half-precision (e.g. torch.float16)
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
This is the configuration class to store the configuration of a . 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.
Configuration objects inherit from and can be used to control the model outputs. Read the documentation from for more information.
config () — 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 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 . 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
Indices can be obtained using . See and for details.
Indices can be obtained using . See and for details.
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 for more information on the default strategy.
return_dict (bool, optional) — Whether or not to return a instead of a plain tuple.
The forward method, overrides the __call__ special method.
( input_ids: LongTensor = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: typing.Optional[typing.List[torch.FloatTensor]] = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Nonelabels: typing.Optional[torch.LongTensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None ) → or tuple(torch.FloatTensor)
Indices can be obtained using . See and for details.
Indices can be obtained using . See and for details.
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 for more information on the default strategy.
return_dict (bool, optional) — Whether or not to return a instead of a plain tuple.
or tuple(torch.FloatTensor)
A 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 () and inputs.
The forward method, overrides the __call__ special method.
config () — 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 method to load the model weights.
uses the last token in order to do the classification, as other causal models (e.g. GPT-2) do.
This model inherits from . 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
Indices can be obtained using . See and for details.
Indices can be obtained using . See and for details.
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 for more information on the default strategy.
return_dict (bool, optional) — Whether or not to return a instead of a plain tuple.
The forward method, overrides the __call__ special method.
