model: support qwen3-vl series (#10323)
Co-authored-by: ocss884 <ocss.lin@gmail.com> Co-authored-by: cao1zhg <653506626@qq.com> Co-authored-by: yhyang201 <yhyang201@gmail.com> Co-authored-by: yhyang201 <47235274+yhyang201@users.noreply.github.com> Co-authored-by: 瑀澈 <yuche.lz@alibaba-inc.com> Co-authored-by: Mick <mickjagger19@icloud.com> Co-authored-by: Yineng Zhang <me@zhyncs.com>
This commit is contained in:
Zheng Li
@@ -749,6 +749,8 @@ multimodal_model_archs = [
|
||||
"Qwen2AudioForConditionalGeneration",
|
||||
"Qwen2VLForConditionalGeneration",
|
||||
"Qwen2_5_VLForConditionalGeneration",
|
||||
"Qwen3VLForConditionalGeneration",
|
||||
"Qwen3VLMoeForConditionalGeneration",
|
||||
"KimiVLForConditionalGeneration",
|
||||
"InternVLChatModel",
|
||||
"InternS1ForConditionalGeneration",
|
||||
|
||||
586
python/sglang/srt/configs/qwen3_vl.py
Normal file
586
python/sglang/srt/configs/qwen3_vl.py
Normal file
@@ -0,0 +1,586 @@
|
||||
from typing import Optional, Union
|
||||
|
||||
from transformers import PretrainedConfig
|
||||
from transformers.modeling_rope_utils import rope_config_validation
|
||||
|
||||
|
||||
class Qwen3VLVisionConfig(PretrainedConfig):
|
||||
model_type = "qwen3_vl"
|
||||
base_config_key = "vision_config"
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
depth=27,
|
||||
hidden_size=1152,
|
||||
hidden_act="gelu_pytorch_tanh",
|
||||
intermediate_size=4304,
|
||||
num_heads=16,
|
||||
in_channels=3,
|
||||
patch_size=16,
|
||||
spatial_merge_size=2,
|
||||
temporal_patch_size=2,
|
||||
out_hidden_size=3584,
|
||||
num_position_embeddings=2304,
|
||||
deepstack_visual_indexes=[8, 16, 24],
|
||||
initializer_range=0.02,
|
||||
**kwargs,
|
||||
):
|
||||
super().__init__(**kwargs)
|
||||
|
||||
self.depth = depth
|
||||
self.hidden_size = hidden_size
|
||||
self.hidden_act = hidden_act
|
||||
self.intermediate_size = intermediate_size
|
||||
self.num_heads = num_heads
|
||||
self.in_channels = in_channels
|
||||
self.patch_size = patch_size
|
||||
self.spatial_merge_size = spatial_merge_size
|
||||
self.temporal_patch_size = temporal_patch_size
|
||||
self.out_hidden_size = out_hidden_size
|
||||
self.num_position_embeddings = num_position_embeddings
|
||||
self.initializer_range = initializer_range
|
||||
self.deepstack_visual_indexes = deepstack_visual_indexes
|
||||
|
||||
|
||||
class Qwen3VLTextConfig(PretrainedConfig):
|
||||
r"""
|
||||
This is the configuration class to store the configuration of a [`Qwen3VLTextModel`]. It is used to instantiate a
|
||||
Qwen3-VL model according to the specified arguments, defining the model architecture. Instantiating a configuration
|
||||
with the defaults will yield a similar configuration to that of
|
||||
Qwen3-VL-4B-Instruct [Qwen/Qwen3-VL-4B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct).
|
||||
|
||||
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||||
documentation from [`PretrainedConfig`] for more information.
|
||||
|
||||
Args:
|
||||
vocab_size (`int`, *optional*, defaults to 151936):
|
||||
Vocabulary size of the Qwen3VL model. Defines the number of different tokens that can be represented by the
|
||||
`inputs_ids` passed when calling [`Qwen3VLModel`]
|
||||
hidden_size (`int`, *optional*, defaults to 4096):
|
||||
Dimension of the hidden representations.
|
||||
intermediate_size (`int`, *optional*, defaults to 22016):
|
||||
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 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, check out [this
|
||||
paper](https://huggingface.co/papers/2305.13245). If it is not specified, will default to `32`.
|
||||
head_dim (`int`, *optional*, defaults to 128):
|
||||
The dimension of the head. If not specified, will default to `hidden_size // num_attention_heads`.
|
||||
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 128000):
|
||||
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`.
|
||||
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 5000000.0):
|
||||
The base period of the RoPE embeddings.
|
||||
rope_scaling (`Dict`, *optional*):
|
||||
Dictionary containing the scaling configuration for the RoPE embeddings. NOTE: if you apply new rope type
|
||||
and you expect the model to work on longer `max_position_embeddings`, we recommend you to update this value
|
||||
accordingly.
|
||||
Expected contents:
|
||||
`rope_type` (`str`):
|
||||
The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope',
|
||||
'llama3'], with 'default' being the original RoPE implementation.
|
||||
`factor` (`float`, *optional*):
|
||||
Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In
|
||||
most scaling types, a `factor` of x will enable the model to handle sequences of length x *
|
||||
original maximum pre-trained length.
|
||||
`original_max_position_embeddings` (`int`, *optional*):
|
||||
Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during
|
||||
pretraining.
|
||||
`attention_factor` (`float`, *optional*):
|
||||
Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention
|
||||
computation. If unspecified, it defaults to value recommended by the implementation, using the
|
||||
`factor` field to infer the suggested value.
|
||||
`beta_fast` (`float`, *optional*):
|
||||
Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear
|
||||
ramp function. If unspecified, it defaults to 32.
|
||||
`beta_slow` (`float`, *optional*):
|
||||
Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear
|
||||
ramp function. If unspecified, it defaults to 1.
|
||||
`short_factor` (`list[float]`, *optional*):
|
||||
Only used with 'longrope'. The scaling factor to be applied to short contexts (<
|
||||
`original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
|
||||
size divided by the number of attention heads divided by 2
|
||||
`long_factor` (`list[float]`, *optional*):
|
||||
Only used with 'longrope'. The scaling factor to be applied to long contexts (<
|
||||
`original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
|
||||
size divided by the number of attention heads divided by 2
|
||||
`low_freq_factor` (`float`, *optional*):
|
||||
Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE
|
||||
`high_freq_factor` (`float`, *optional*):
|
||||
Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE
|
||||
attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
|
||||
Whether to use a bias in the query, key, value and output projection layers during self-attention.
|
||||
attention_dropout (`float`, *optional*, defaults to 0.0):
|
||||
The dropout ratio for the attention probabilities.
|
||||
|
||||
```python
|
||||
>>> from transformers import Qwen3VLTextModel, Qwen3VLTextConfig
|
||||
|
||||
>>> # Initializing a Qwen3VL style configuration
|
||||
>>> configuration = Qwen3VLTextConfig()
|
||||
|
||||
>>> # Initializing a model from the Qwen3-VL-7B style configuration
|
||||
>>> model = Qwen3VLTextModel(configuration)
|
||||
|
||||
>>> # Accessing the model configuration
|
||||
>>> configuration = model.config
|
||||
```"""
|
||||
|
||||
model_type = "qwen3_vl_text"
|
||||
base_config_key = "text_config"
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
vocab_size=151936,
|
||||
hidden_size=4096,
|
||||
intermediate_size=22016,
|
||||
num_hidden_layers=32,
|
||||
num_attention_heads=32,
|
||||
num_key_value_heads=32,
|
||||
head_dim=128,
|
||||
hidden_act="silu",
|
||||
max_position_embeddings=128000,
|
||||
initializer_range=0.02,
|
||||
rms_norm_eps=1e-6,
|
||||
use_cache=True,
|
||||
tie_word_embeddings=False,
|
||||
rope_theta=5000000.0,
|
||||
rope_scaling=None,
|
||||
attention_bias=False,
|
||||
attention_dropout=0.0,
|
||||
**kwargs,
|
||||
):
|
||||
self.vocab_size = vocab_size
|
||||
self.max_position_embeddings = max_position_embeddings
|
||||
self.hidden_size = hidden_size
|
||||
self.intermediate_size = intermediate_size
|
||||
self.num_hidden_layers = num_hidden_layers
|
||||
self.num_attention_heads = num_attention_heads
|
||||
|
||||
# for backward compatibility
|
||||
if num_key_value_heads is None:
|
||||
num_key_value_heads = num_attention_heads
|
||||
|
||||
self.num_key_value_heads = num_key_value_heads
|
||||
self.head_dim = head_dim
|
||||
self.hidden_act = hidden_act
|
||||
self.initializer_range = initializer_range
|
||||
self.rms_norm_eps = rms_norm_eps
|
||||
self.use_cache = use_cache
|
||||
self.rope_theta = rope_theta
|
||||
self.rope_scaling = rope_scaling
|
||||
self.attention_bias = attention_bias
|
||||
self.attention_dropout = attention_dropout
|
||||
|
||||
rope_config_validation(self, ignore_keys={"mrope_section", "mrope_interleaved"})
|
||||
|
||||
super().__init__(tie_word_embeddings=tie_word_embeddings, **kwargs)
|
||||
|
||||
|
||||
class Qwen3VLConfig(PretrainedConfig):
|
||||
r"""
|
||||
This is the configuration class to store the configuration of a [`Qwen3VLModel`]. It is used to instantiate a
|
||||
Qwen3-VL model according to the specified arguments, defining the model architecture. Instantiating a configuration
|
||||
with the defaults will yield a similar configuration to that of
|
||||
Qwen3-VL-4B-Instruct [Qwen/Qwen3-VL-4B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct).
|
||||
|
||||
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||||
documentation from [`PretrainedConfig`] for more information.
|
||||
|
||||
|
||||
Args:
|
||||
text_config (`Union[PreTrainedConfig, dict]`, *optional*, defaults to `Qwen3VLTextConfig`):
|
||||
The config object or dictionary of the text backbone.
|
||||
vision_config (`Union[PreTrainedConfig, dict]`, *optional*, defaults to `Qwen3VLVisionConfig`):
|
||||
The config object or dictionary of the vision backbone.
|
||||
image_token_id (`int`, *optional*, defaults to 151655):
|
||||
The image token index to encode the image prompt.
|
||||
video_token_id (`int`, *optional*, defaults to 151656):
|
||||
The video token index to encode the image prompt.
|
||||
vision_start_token_id (`int`, *optional*, defaults to 151652):
|
||||
The start token index to encode the image prompt.
|
||||
vision_end_token_id (`int`, *optional*, defaults to 151653):
|
||||
The end token index to encode the image prompt.
|
||||
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
||||
Whether to tie the word embeddings.
|
||||
|
||||
```python
|
||||
>>> from transformers import Qwen3VLForConditionalGeneration, Qwen3VLConfig
|
||||
|
||||
>>> # Initializing a Qwen3-VL style configuration
|
||||
>>> configuration = Qwen3VLConfig()
|
||||
|
||||
>>> # Initializing a model from the Qwen3-VL-4B style configuration
|
||||
>>> model = Qwen3VLForConditionalGeneration(configuration)
|
||||
|
||||
>>> # Accessing the model configuration
|
||||
>>> configuration = model.config
|
||||
```"""
|
||||
|
||||
model_type = "qwen3_vl"
|
||||
sub_configs = {
|
||||
"vision_config": Qwen3VLVisionConfig,
|
||||
"text_config": Qwen3VLTextConfig,
|
||||
}
|
||||
keys_to_ignore_at_inference = ["past_key_values"]
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
text_config=None,
|
||||
vision_config=None,
|
||||
image_token_id=151655,
|
||||
video_token_id=151656,
|
||||
vision_start_token_id=151652,
|
||||
vision_end_token_id=151653,
|
||||
tie_word_embeddings=False,
|
||||
**kwargs,
|
||||
):
|
||||
if isinstance(vision_config, dict):
|
||||
self.vision_config = self.sub_configs["vision_config"](**vision_config)
|
||||
elif vision_config is None:
|
||||
self.vision_config = self.sub_configs["vision_config"]()
|
||||
|
||||
if isinstance(text_config, dict):
|
||||
self.text_config = self.sub_configs["text_config"](**text_config)
|
||||
elif text_config is None:
|
||||
self.text_config = self.sub_configs["text_config"]()
|
||||
|
||||
self.image_token_id = image_token_id
|
||||
self.video_token_id = video_token_id
|
||||
self.vision_start_token_id = vision_start_token_id
|
||||
self.vision_end_token_id = vision_end_token_id
|
||||
super().__init__(**kwargs, tie_word_embeddings=tie_word_embeddings)
|
||||
|
||||
|
||||
class Qwen3VLMoeTextConfig(PretrainedConfig):
|
||||
r"""
|
||||
This is the configuration class to store the configuration of a [`Qwen3VLMoeTextModel`]. It is used to instantiate a
|
||||
Qwen3-VL-MOE model according to the specified arguments, defining the model architecture. Instantiating a configuration
|
||||
with the defaults will yield a similar configuration to that of
|
||||
Qwen3-VL-30B-A3B-Instruct [Qwen/Qwen3-VL-30B-A3B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct).
|
||||
|
||||
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||||
documentation from [`PretrainedConfig`] for more information.
|
||||
|
||||
Args:
|
||||
vocab_size (`int`, *optional*, defaults to 151936):
|
||||
Vocabulary size of the Qwen2MoE model. Defines the number of different tokens that can be represented by the
|
||||
`inputs_ids` passed when calling [`Qwen2MoeModel`]
|
||||
hidden_size (`int`, *optional*, defaults to 2048):
|
||||
Dimension of the hidden representations.
|
||||
intermediate_size (`int`, *optional*, defaults to 5632):
|
||||
Dimension of the MLP representations.
|
||||
num_hidden_layers (`int`, *optional*, defaults to 24):
|
||||
Number of hidden layers in the Transformer encoder.
|
||||
num_attention_heads (`int`, *optional*, defaults to 16):
|
||||
Number of attention heads for each attention layer in the Transformer encoder.
|
||||
num_key_value_heads (`int`, *optional*, defaults to 16):
|
||||
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`.
|
||||
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 128000):
|
||||
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`.
|
||||
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 5000000.0):
|
||||
The base period of the RoPE embeddings.
|
||||
attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
|
||||
Whether to use a bias in the query, key, value and output projection layers during self-attention.
|
||||
attention_dropout (`float`, *optional*, defaults to 0.0):
|
||||
The dropout ratio for the attention probabilities.
|
||||
decoder_sparse_step (`int`, *optional*, defaults to 1):
|
||||
The frequency of the MoE layer.
|
||||
moe_intermediate_size (`int`, *optional*, defaults to 1408):
|
||||
Intermediate size of the routed expert.
|
||||
num_experts_per_tok (`int`, *optional*, defaults to 4):
|
||||
Number of selected experts.
|
||||
num_experts (`int`, *optional*, defaults to 60):
|
||||
Number of routed experts.
|
||||
norm_topk_prob (`bool`, *optional*, defaults to `True`):
|
||||
Whether to normalize the topk probabilities.
|
||||
mlp_only_layers (`List[int]`, *optional*, defaults to `[]`):
|
||||
Indicate which layers use Qwen3VLMoeMLP rather than Qwen3VLMoeSparseMoeBlock
|
||||
The list contains layer index, from 0 to num_layers-1 if we have num_layers layers
|
||||
If `mlp_only_layers` is empty, `decoder_sparse_step` is used to determine the sparsity.
|
||||
rope_scaling (`Dict`, *optional*):
|
||||
Dictionary containing the scaling configuration for the RoPE embeddings. NOTE: if you apply new rope type
|
||||
and you expect the model to work on longer `max_position_embeddings`, we recommend you to update this value
|
||||
accordingly.
|
||||
Expected contents:
|
||||
`rope_type` (`str`):
|
||||
The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope',
|
||||
'llama3'], with 'default' being the original RoPE implementation.
|
||||
`factor` (`float`, *optional*):
|
||||
Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In
|
||||
most scaling types, a `factor` of x will enable the model to handle sequences of length x *
|
||||
original maximum pre-trained length.
|
||||
`original_max_position_embeddings` (`int`, *optional*):
|
||||
Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during
|
||||
pretraining.
|
||||
`attention_factor` (`float`, *optional*):
|
||||
Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention
|
||||
computation. If unspecified, it defaults to value recommended by the implementation, using the
|
||||
`factor` field to infer the suggested value.
|
||||
`beta_fast` (`float`, *optional*):
|
||||
Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear
|
||||
ramp function. If unspecified, it defaults to 32.
|
||||
`beta_slow` (`float`, *optional*):
|
||||
Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear
|
||||
ramp function. If unspecified, it defaults to 1.
|
||||
`short_factor` (`List[float]`, *optional*):
|
||||
Only used with 'longrope'. The scaling factor to be applied to short contexts (<
|
||||
`original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
|
||||
size divided by the number of attention heads divided by 2
|
||||
`long_factor` (`List[float]`, *optional*):
|
||||
Only used with 'longrope'. The scaling factor to be applied to long contexts (<
|
||||
`original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
|
||||
size divided by the number of attention heads divided by 2
|
||||
`low_freq_factor` (`float`, *optional*):
|
||||
Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE
|
||||
`high_freq_factor` (`float`, *optional*):
|
||||
Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE
|
||||
head_dim (`int`, *optional*):
|
||||
The dimension of the head. If not specified, will default to `hidden_size // num_attention_heads`.
|
||||
|
||||
```python
|
||||
>>> from transformers import Qwen3VLMoeForConditionalGeneration, Qwen3VLMoeConfig
|
||||
|
||||
>>> # Initializing a Qwen3VLMoe style configuration
|
||||
>>> configuration = Qwen3VLMoeConfig()
|
||||
|
||||
>>> # Initializing a model from the Qwen3-VL-30B-A3B style configuration
|
||||
>>> model = Qwen3VLMoeForConditionalGeneration(configuration)
|
||||
|
||||
>>> # Accessing the model configuration
|
||||
>>> configuration = model.config
|
||||
```"""
|
||||
|
||||
model_type = "qwen3_vl_moe_text"
|
||||
base_config_key = "text_config"
|
||||
keys_to_ignore_at_inference = ["past_key_values"]
|
||||
# Default tensor parallel plan for base model `Qwen3VLMoe`
|
||||
base_model_tp_plan = {
|
||||
"layers.*.self_attn.q_proj": "colwise",
|
||||
"layers.*.self_attn.k_proj": "colwise",
|
||||
"layers.*.self_attn.v_proj": "colwise",
|
||||
"layers.*.self_attn.o_proj": "rowwise",
|
||||
"layers.*.mlp.gate_proj": "colwise",
|
||||
"layers.*.mlp.up_proj": "colwise",
|
||||
"layers.*.mlp.down_proj": "rowwise",
|
||||
}
|
||||
base_model_pp_plan = {
|
||||
"embed_tokens": (["input_ids"], ["inputs_embeds"]),
|
||||
"layers": (["hidden_states", "attention_mask"], ["hidden_states"]),
|
||||
"norm": (["hidden_states"], ["hidden_states"]),
|
||||
}
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
vocab_size=151936,
|
||||
hidden_size=2048,
|
||||
intermediate_size=5632,
|
||||
num_hidden_layers=24,
|
||||
num_attention_heads=16,
|
||||
num_key_value_heads=16,
|
||||
hidden_act="silu",
|
||||
max_position_embeddings=128000,
|
||||
initializer_range=0.02,
|
||||
rms_norm_eps=1e-6,
|
||||
use_cache=True,
|
||||
tie_word_embeddings=False,
|
||||
rope_theta=5000000.0,
|
||||
attention_bias=False,
|
||||
attention_dropout=0.0,
|
||||
decoder_sparse_step=1,
|
||||
moe_intermediate_size=1408,
|
||||
num_experts_per_tok=4,
|
||||
num_experts=60,
|
||||
norm_topk_prob=True,
|
||||
mlp_only_layers=None,
|
||||
rope_scaling=None,
|
||||
head_dim=None,
|
||||
**kwargs,
|
||||
):
|
||||
self.vocab_size = vocab_size
|
||||
self.max_position_embeddings = max_position_embeddings
|
||||
self.hidden_size = hidden_size
|
||||
self.intermediate_size = intermediate_size
|
||||
self.num_hidden_layers = num_hidden_layers
|
||||
self.num_attention_heads = num_attention_heads
|
||||
|
||||
# for backward compatibility
|
||||
if num_key_value_heads is None:
|
||||
num_key_value_heads = num_attention_heads
|
||||
|
||||
self.num_key_value_heads = num_key_value_heads
|
||||
self.hidden_act = hidden_act
|
||||
self.initializer_range = initializer_range
|
||||
self.rms_norm_eps = rms_norm_eps
|
||||
self.use_cache = use_cache
|
||||
self.rope_theta = rope_theta
|
||||
self.attention_bias = attention_bias
|
||||
self.attention_dropout = attention_dropout
|
||||
self.rope_scaling = rope_scaling
|
||||
self.head_dim = head_dim or hidden_size // num_attention_heads
|
||||
|
||||
rope_config_validation(self, ignore_keys={"mrope_section", "mrope_interleaved"})
|
||||
|
||||
# MoE arguments
|
||||
self.decoder_sparse_step = decoder_sparse_step
|
||||
self.moe_intermediate_size = moe_intermediate_size
|
||||
self.num_experts_per_tok = num_experts_per_tok
|
||||
self.num_experts = num_experts
|
||||
self.norm_topk_prob = norm_topk_prob
|
||||
self.mlp_only_layers = [] if mlp_only_layers is None else mlp_only_layers
|
||||
|
||||
super().__init__(tie_word_embeddings=tie_word_embeddings, **kwargs)
|
||||
|
||||
|
||||
class Qwen3VLMoeVisionConfig(PretrainedConfig):
|
||||
model_type = "qwen3_vl_moe"
|
||||
base_config_key = "vision_config"
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
depth=27,
|
||||
hidden_size=1152,
|
||||
hidden_act="gelu_pytorch_tanh",
|
||||
intermediate_size=4304,
|
||||
num_heads=16,
|
||||
in_channels=3,
|
||||
patch_size=16,
|
||||
spatial_merge_size=2,
|
||||
temporal_patch_size=2,
|
||||
out_hidden_size=3584,
|
||||
num_position_embeddings=2304,
|
||||
deepstack_visual_indexes=[8, 16, 24],
|
||||
initializer_range=0.02,
|
||||
**kwargs,
|
||||
):
|
||||
super().__init__(**kwargs)
|
||||
|
||||
self.depth = depth
|
||||
self.hidden_size = hidden_size
|
||||
self.hidden_act = hidden_act
|
||||
self.intermediate_size = intermediate_size
|
||||
self.num_heads = num_heads
|
||||
self.in_channels = in_channels
|
||||
self.patch_size = patch_size
|
||||
self.spatial_merge_size = spatial_merge_size
|
||||
self.temporal_patch_size = temporal_patch_size
|
||||
self.out_hidden_size = out_hidden_size
|
||||
self.num_position_embeddings = num_position_embeddings
|
||||
self.initializer_range = initializer_range
|
||||
self.deepstack_visual_indexes = deepstack_visual_indexes
|
||||
|
||||
|
||||
class Qwen3VLMoeConfig(PretrainedConfig):
|
||||
r"""
|
||||
This is the configuration class to store the configuration of a [`Qwen3VLMoeModel`]. It is used to instantiate a
|
||||
Qwen3-VL-MOE model according to the specified arguments, defining the model architecture. Instantiating a configuration
|
||||
with the defaults will yield a similar configuration to that of
|
||||
Qwen3-VL-30B-A3B-Instruct [Qwen/Qwen3-VL-30B-A3B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct).
|
||||
|
||||
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||||
documentation from [`PretrainedConfig`] for more information.
|
||||
|
||||
|
||||
Args:
|
||||
text_config (`Union[PreTrainedConfig, dict]`, *optional*, defaults to `Qwen3VLMoeTextConfig`):
|
||||
The config object or dictionary of the text backbone.
|
||||
vision_config (`Union[PreTrainedConfig, dict]`, *optional*, defaults to `Qwen3VLMoeVisionConfig`):
|
||||
The config object or dictionary of the vision backbone.
|
||||
image_token_id (`int`, *optional*, defaults to 151655):
|
||||
The image token index to encode the image prompt.
|
||||
video_token_id (`int`, *optional*, defaults to 151656):
|
||||
The video token index to encode the image prompt.
|
||||
vision_start_token_id (`int`, *optional*, defaults to 151652):
|
||||
The start token index to encode the image prompt.
|
||||
vision_end_token_id (`int`, *optional*, defaults to 151653):
|
||||
The end token index to encode the image prompt.
|
||||
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
||||
Whether to tie the word embeddings.
|
||||
|
||||
```python
|
||||
>>> from transformers import Qwen3VLMoeForConditionalGeneration, Qwen3VLMoeConfig
|
||||
|
||||
>>> # Initializing a Qwen3-VL-MOE style configuration
|
||||
>>> configuration = Qwen3VLMoeConfig()
|
||||
|
||||
>>> # Initializing a model from the Qwen3-VL-30B-A3B style configuration
|
||||
>>> model = Qwen3VLMoeForConditionalGeneration(configuration)
|
||||
|
||||
>>> # Accessing the model configuration
|
||||
>>> configuration = model.config
|
||||
```"""
|
||||
|
||||
model_type = "qwen3_vl_moe"
|
||||
sub_configs = {
|
||||
"vision_config": Qwen3VLMoeVisionConfig,
|
||||
"text_config": Qwen3VLMoeTextConfig,
|
||||
}
|
||||
keys_to_ignore_at_inference = ["past_key_values"]
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
text_config=None,
|
||||
vision_config=None,
|
||||
image_token_id=151655,
|
||||
video_token_id=151656,
|
||||
vision_start_token_id=151652,
|
||||
vision_end_token_id=151653,
|
||||
tie_word_embeddings=False,
|
||||
**kwargs,
|
||||
):
|
||||
if isinstance(vision_config, dict):
|
||||
self.vision_config = self.sub_configs["vision_config"](**vision_config)
|
||||
elif vision_config is None:
|
||||
self.vision_config = self.sub_configs["vision_config"]()
|
||||
|
||||
if isinstance(text_config, dict):
|
||||
self.text_config = self.sub_configs["text_config"](**text_config)
|
||||
elif text_config is None:
|
||||
self.text_config = self.sub_configs["text_config"]()
|
||||
|
||||
self.image_token_id = image_token_id
|
||||
self.video_token_id = video_token_id
|
||||
self.vision_start_token_id = vision_start_token_id
|
||||
self.vision_end_token_id = vision_end_token_id
|
||||
super().__init__(**kwargs, tie_word_embeddings=tie_word_embeddings)
|
||||
|
||||
|
||||
__all__ = [
|
||||
"Qwen3VLMoeConfig",
|
||||
"Qwen3VLMoeVisionConfig",
|
||||
"Qwen3VLConfig",
|
||||
"Qwen3VLVisionConfig",
|
||||
]
|
||||
Reference in New Issue
Block a user