Qwen2 is the new model series of large language models from the Qwen team. Previously, we released the Qwen series, including Qwen-72B, Qwen-1.8B, Qwen-VL, Qwen-Audio, etc.
Qwen2 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes.
Qwen2-7B-beta and Qwen2-7B-Chat-beta can be found on the Huggingface Hub
In the following, we demonstrate how to use Qwen2-7B-Chat-beta for the inference. Note that we have used the ChatML format for dialog, in this demo we show how to leverage apply_chat_template for this purpose.
>>> from transformers import AutoModelForCausalLM, AutoTokenizer
>>> device = "cuda" # the device to load the model onto
>>> model = AutoModelForCausalLM.from_pretrained("Qwen2/Qwen2-7B-Chat-beta", device_map="auto")
>>> tokenizer = AutoTokenizer.from_pretrained("Qwen2/Qwen2-7B-Chat-beta")
>>> prompt = "Give me a short introduction to large language model."
>>> messages = [{"role": "user", "content": prompt}]
>>> text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
>>> model_inputs = tokenizer([text], return_tensors="pt").to(device)
>>> generated_ids = model.generate(model_inputs.input_ids, max_new_tokens=512, do_sample=True)
>>> generated_ids = [output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)]
>>> response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]( vocab_size = 151936 hidden_size = 4096 intermediate_size = 22016 num_hidden_layers = 32 num_attention_heads = 32 num_key_value_heads = 32 hidden_act = 'silu' max_position_embeddings = 32768 initializer_range = 0.02 rms_norm_eps = 1e-06 use_cache = True tie_word_embeddings = False rope_theta = 10000.0 use_sliding_window = False sliding_window = 4096 max_window_layers = 28 attention_dropout = 0.0 **kwargs )
Parameters
int, optional, defaults to 151936) —
Vocabulary size of the Qwen2 model. Defines the number of different tokens that can be represented by the
inputs_ids passed when calling Qwen2Model int, optional, defaults to 4096) —
Dimension of the hidden representations. int, optional, defaults to 22016) —
Dimension of the MLP representations. int, optional, defaults to 32) —
Number of hidden layers in the Transformer encoder. int, optional, defaults to 32) —
Number of attention heads for each attention layer in the Transformer encoder. int, optional, defaults to 32) —
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 32`. str or function, optional, defaults to "silu") —
The non-linear activation function (function or string) in the decoder. int, optional, defaults to 32768) —
The maximum sequence length that this model might ever be used with. float, optional, defaults to 0.02) —
The standard deviation of the truncated_normal_initializer for initializing all weight matrices. float, optional, defaults to 1e-06) —
The epsilon used by the rms normalization layers. 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. bool, optional, defaults to False) —
Whether the model’s input and output word embeddings should be tied. float, optional, defaults to 10000.0) —
The base period of the RoPE embeddings. bool, optional, defaults to False) —
Whether to use sliding window attention. int, optional, defaults to 4096) —
Sliding window attention (SWA) window size. If not specified, will default to 4096. int, optional, defaults to 28) —
The number of layers that use SWA (Sliding Window Attention). The bottom layers use SWA while the top use full attention. float, optional, defaults to 0.0) —
The dropout ratio for the attention probabilities. This is the configuration class to store the configuration of a Qwen2Model. It is used to instantiate a Qwen2 model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of Qwen2-7B-beta Qwen/Qwen2-7B-beta.
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 Qwen2Model, Qwen2Config
>>> # Initializing a Qwen2 style configuration
>>> configuration = Qwen2Config()
>>> # Initializing a model from the Qwen2-7B style configuration
>>> model = Qwen2Model(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config( vocab_file merges_file errors = 'replace' unk_token = '<|endoftext|>' bos_token = None eos_token = '<|endoftext|>' pad_token = '<|endoftext|>' clean_up_tokenization_spaces = False split_special_tokens = False **kwargs )
Parameters
str) —
Path to the vocabulary file. str) —
Path to the merges file. str, optional, defaults to "replace") —
Paradigm to follow when decoding bytes to UTF-8. See
bytes.decode for more information. str, optional, defaults to "<|endoftext|>") —
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. str, optional) —
The beginning of sequence token. Not applicable for this tokenizer. str, optional, defaults to "<|endoftext|>") —
The end of sequence token. str, optional, defaults to "<|endoftext|>") —
The token used for padding, for example when batching sequences of different lengths. bool, optional, defaults to False) —
Whether or not the model should cleanup the spaces that were added when splitting the input text during the
tokenization process. Not applicable to this tokenizer, since tokenization does not add spaces. bool, optional, defaults to False) —
Whether or not the special tokens should be split during the tokenization process. The default behavior is
to not split special tokens. This means that if <|endoftext|> is the eos_token, then tokenizer.tokenize("<|endoftext|>") = ['<|endoftext|>]. Otherwise, if split_special_tokens=True, then tokenizer.tokenize("<|endoftext|>") will be give ['<', '|', 'endo', 'ft', 'ext', '|', '>']. This argument is only supported for slow tokenizers for the moment. Construct a Qwen2 tokenizer. Based on byte-level Byte-Pair-Encoding.
Same with GPT2Tokenizer, this tokenizer has been trained to treat spaces like parts of the tokens so a word will
be encoded differently whether it is at the beginning of the sentence (without space) or not:
>>> from transformers import Qwen2Tokenizer
>>> tokenizer = Qwen2Tokenizer.from_pretrained("Qwen/Qwen-tokenizer")
>>> tokenizer("Hello world")["input_ids"]
[9707, 1879]
>>> tokenizer(" Hello world")["input_ids"]
[21927, 1879]You should not use GPT2Tokenizer instead, because of the different pretokenization rules.
This tokenizer inherits from PreTrainedTokenizer which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.
( vocab_file = None merges_file = None tokenizer_file = None unk_token = '<|endoftext|>' bos_token = None eos_token = '<|endoftext|>' pad_token = '<|endoftext|>' **kwargs )
Parameters
str, optional) —
Path to the vocabulary file. str, optional) —
Path to the merges file. str, optional) —
Path to tokenizers file (generally has a .json extension) that
contains everything needed to load the tokenizer. str, optional, defaults to "<|endoftext|>") —
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. Not applicable to this tokenizer. str, optional) —
The beginning of sequence token. Not applicable for this tokenizer. str, optional, defaults to "<|endoftext|>") —
The end of sequence token. str, optional, defaults to "<|endoftext|>") —
The token used for padding, for example when batching sequences of different lengths. Construct a “fast” Qwen2 tokenizer (backed by HuggingFace’s tokenizers library). Based on byte-level Byte-Pair-Encoding.
Same with GPT2Tokenizer, this tokenizer has been trained to treat spaces like parts of the tokens so a word will
be encoded differently whether it is at the beginning of the sentence (without space) or not:
>>> from transformers import Qwen2TokenizerFast
>>> tokenizer = Qwen2TokenizerFast.from_pretrained("Qwen/Qwen-tokenizer")
>>> tokenizer("Hello world")["input_ids"]
[9707, 1879]
>>> tokenizer(" Hello world")["input_ids"]
[21927, 1879]This tokenizer inherits from PreTrainedTokenizerFast which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.
( config: Qwen2Config )
Parameters
The bare Qwen2 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 Qwen2DecoderLayer
( 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
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.
torch.Tensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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.
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].
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:
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).
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. 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). bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under returned
tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. The Qwen2Model 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.
( 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
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.
torch.Tensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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.
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].
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:
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).
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. 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). bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under returned
tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. 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 (Qwen2Config) 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 Qwen2ForCausalLM 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.
Example:
>>> from transformers import AutoTokenizer, Qwen2ForCausalLM
>>> model = Qwen2ForCausalLM.from_pretrained(PATH_TO_CONVERTED_WEIGHTS)
>>> tokenizer = AutoTokenizer.from_pretrained(PATH_TO_CONVERTED_TOKENIZER)
>>> 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."( config )
Parameters
The Qwen2 Model transformer with a sequence classification head on top (linear layer).
Qwen2ForSequenceClassification 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.
( 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 )
Parameters
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.
torch.Tensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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.
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].
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:
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).
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. 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). bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under returned
tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. 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 Qwen2ForSequenceClassification 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.