You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
451 lines
22 KiB
451 lines
22 KiB
# Copyright 2025 The LLAMA4 and HuggingFace Inc. team. All rights reserved.
|
|
#
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
|
|
from ...configuration_utils import PreTrainedConfig, layer_type_validation
|
|
from ...modeling_rope_utils import RopeParameters
|
|
from ...utils import logging
|
|
|
|
|
|
logger = logging.get_logger(__name__)
|
|
|
|
|
|
class Llama4VisionConfig(PreTrainedConfig):
|
|
r"""
|
|
This is the configuration class to store the configuration of a [`Llama4VisionModel`]. It is used to instantiate a
|
|
Llama4 vision 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 Llama4 109B.
|
|
|
|
e.g. [meta-llama/Llama-4-Scout-17B-16E](https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E)
|
|
|
|
Configuration objects inherit from [`PreTrainedConfig`] and can be used to control the model outputs. Read the
|
|
documentation from [`PreTrainedConfig`] for more information.
|
|
|
|
Args:
|
|
hidden_size (`int`, *optional*, defaults to 768):
|
|
Dimensionality of the encoder layers and the pooler layer.
|
|
hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`):
|
|
The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
|
|
`"relu"`, `"selu"` and `"gelu_new"` `"quick_gelu"` are supported.
|
|
num_hidden_layers (`int`, *optional*, defaults to 34):
|
|
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_channels (`int`, *optional*, defaults to 3):
|
|
Number of channels in the input image.
|
|
intermediate_size (`int`, *optional*, defaults to 5632):
|
|
Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
|
|
vision_output_dim (`int`, *optional*, defaults to 7680):
|
|
Dimensionality of the vision model output. Includes output of transformer
|
|
encoder with intermediate layers and global transformer encoder.
|
|
image_size (`int`, *optional*, defaults to 448):
|
|
The size (resolution) of each image *tile*.
|
|
patch_size (`int`, *optional*, defaults to 14):
|
|
The size (resolution) of each patch.
|
|
norm_eps (`float`, *optional*, defaults to 1e-05):
|
|
The epsilon used by the layer normalization layers.
|
|
vision_feature_select_strategy (`str`, *optional*, defaults to `"default"`):
|
|
Controls which vision tokens are kept from the backbone. `"default"` drops the CLS token and `"full"` keeps all tokens.
|
|
initializer_range (`float`, *optional*, defaults to 0.02):
|
|
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
|
pixel_shuffle_ratio (`float`, *optional*, defaults to 0.5):
|
|
Pixel-shuffle ratio for downsampling patch tokens. Smaller values produce fewer tokens (more downsampling).
|
|
projector_input_dim (`int`, *optional*, defaults to 4096):
|
|
Width of the vision adapter MLP before pixel shuffle. Larger value increases capacity and compute.
|
|
projector_output_dim (`int`, *optional*, defaults to 4096):
|
|
Output width of the vision adapter. Larger value yields higher-dimensional image features.
|
|
multi_modal_projector_bias (`bool`, *optional*, defaults to `False`):
|
|
Whether to use bias in the multi-modal projector layers.
|
|
projector_dropout (`float`, *optional*, defaults to 0.0):
|
|
Dropout rate inside the vision adapter MLP. Higher value adds more regularization.
|
|
attention_dropout (`float`, *optional*, defaults to 0.0):
|
|
Dropout rate on vision attention probabilities. Higher value adds more regularization.
|
|
rope_parameters (`RopeParameters`, *optional*):
|
|
RoPE Parameters
|
|
"""
|
|
|
|
base_model_tp_plan = {
|
|
"model.layers.*.self_attn.q_proj": "colwise",
|
|
"model.layers.*.self_attn.k_proj": "colwise",
|
|
"model.layers.*.self_attn.v_proj": "colwise",
|
|
"model.layers.*.self_attn.o_proj": "rowwise",
|
|
"vision_adapter.mlp.fc1": "colwise",
|
|
"vision_adapter.mlp.fc2": "rowwise",
|
|
"patch_embedding.linear": "colwise_gather_output",
|
|
}
|
|
model_type = "llama4_vision_model"
|
|
base_config_key = "vision_config"
|
|
|
|
def __init__(
|
|
self,
|
|
hidden_size: int | None = 768,
|
|
hidden_act: str | None = "gelu",
|
|
num_hidden_layers: int | None = 34,
|
|
num_attention_heads: int | None = 16,
|
|
num_channels: int | None = 3,
|
|
intermediate_size: int | None = 5632,
|
|
vision_output_dim: int | None = 7680,
|
|
image_size: int | None = 448,
|
|
patch_size: int | None = 14,
|
|
norm_eps: float | None = 1e-5,
|
|
vision_feature_select_strategy: str | None = "default",
|
|
initializer_range: float | None = 0.02,
|
|
pixel_shuffle_ratio: float | None = 0.5,
|
|
projector_input_dim: int | None = 4096,
|
|
projector_output_dim: int | None = 4096,
|
|
multi_modal_projector_bias: bool | None = False,
|
|
projector_dropout: float | None = 0.0,
|
|
attention_dropout: float | None = 0.0,
|
|
rope_parameters: RopeParameters | dict[str, RopeParameters] | None = None,
|
|
**kwargs,
|
|
):
|
|
self.hidden_size = hidden_size
|
|
self.hidden_act = hidden_act
|
|
self.num_hidden_layers = num_hidden_layers
|
|
self.num_channels = num_channels
|
|
self.intermediate_size = intermediate_size
|
|
self.image_size = image_size
|
|
self.vision_output_dim = vision_output_dim
|
|
self.patch_size = patch_size
|
|
self.norm_eps = norm_eps
|
|
self.num_attention_heads = num_attention_heads
|
|
self.initializer_range = initializer_range
|
|
self.pixel_shuffle_ratio = pixel_shuffle_ratio
|
|
self.projector_input_dim = projector_input_dim
|
|
self.projector_output_dim = projector_output_dim
|
|
self.multi_modal_projector_bias = multi_modal_projector_bias
|
|
self.projector_dropout = projector_dropout
|
|
self.attention_dropout = attention_dropout
|
|
self.vision_feature_select_strategy = vision_feature_select_strategy
|
|
|
|
self.rope_parameters = rope_parameters
|
|
|
|
super().__init__(**kwargs)
|
|
|
|
|
|
class Llama4TextConfig(PreTrainedConfig):
|
|
r"""
|
|
This is the configuration class to store the configuration of a [`Llama4TextModel`]. It is used to instantiate a
|
|
Llama4 text 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 Llama4 109B.
|
|
|
|
e.g. [meta-llama/Llama-4-Scout-17B-16E](https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E)
|
|
|
|
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 202048):
|
|
Vocabulary size of the Llama4 text model. Defines the maximum number of different tokens that can be represented
|
|
by the `inputs_ids` passed when calling [`Llama4TextModel`].
|
|
hidden_size (`int`, *optional*, defaults to 5120):
|
|
Dimensionality of the embeddings and hidden states.
|
|
intermediate_size (`int`, *optional*, defaults to 8192):
|
|
Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
|
|
intermediate_size_mlp (`int`, *optional*, defaults to 16384):
|
|
Intermediate size of dense MLP layers. Larger value increases FFN capacity and compute.
|
|
num_hidden_layers (`int`, *optional*, defaults to 48):
|
|
Number of hidden layers in the Transformer encoder.
|
|
num_attention_heads (`int`, *optional*, defaults to 40):
|
|
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 not
|
|
specified, will default to `num_attention_heads`.
|
|
head_dim (`int`, *optional*, defaults to 128):
|
|
Per-head attention dimension. Larger value increases head width and compute.
|
|
hidden_act (`str` or `Callable`, *optional*, defaults to `"silu"`):
|
|
The non-linear activation function (function or string) in the encoder and pooler.
|
|
max_position_embeddings (`int`, *optional*, defaults to 131072):
|
|
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-05):
|
|
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.
|
|
pad_token_id (`int`, *optional*, defaults to 128004):
|
|
The id of the padding token.
|
|
bos_token_id (`int`, *optional*, defaults to 1):
|
|
The id of the beginning of sentence token.
|
|
eos_token_id (`int`, *optional*, defaults to 2):
|
|
The id of the end of sentence token.
|
|
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
|
Whether to tie weight embeddings
|
|
attention_dropout (`float`, *optional*, defaults to 0.0):
|
|
Dropout rate on vision attention probabilities. Higher value adds more regularization.
|
|
num_experts_per_tok (`int`, *optional*, defaults to 1):
|
|
Top-k experts routed per token. Higher value uses more experts per token and more compute.
|
|
num_local_experts (`int`, *optional*, defaults to 16):
|
|
Number of experts in each MoE layer. Higher value increases capacity and routing choices.
|
|
moe_layers (`list[int]`, *optional*):
|
|
List of layer indices that use MoE. Overrides `interleave_moe_layer_step` when set.
|
|
interleave_moe_layer_step (`int`, *optional*, defaults to 1):
|
|
Spacing between MoE layers when `moe_layers` is `None`. Larger value means fewer MoE layers.
|
|
use_qk_norm (`bool`, *optional*, defaults to `True`):
|
|
Whether to L2-normalize queries/keys on RoPE layers. Can stabilize attention when enabled.
|
|
output_router_logits (`bool`, *optional*, defaults to `False`):
|
|
Whether to return router logits (and auxiliary loss) in outputs.
|
|
router_aux_loss_coef (`float`, *optional*, defaults to 0.001):
|
|
Weight for the router auxiliary loss. Higher value makes routing loss contribute more to total loss.
|
|
router_jitter_noise (`float`, *optional*, defaults to 0.0):
|
|
Amount of noise added to router logits during training. Higher value increases exploration.
|
|
rope_parameters (`RopeParameters`, *optional*):
|
|
Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain
|
|
a value for `rope_theta` and optionally parameters used for scaling in case you want to use RoPE
|
|
with longer `max_position_embeddings`.
|
|
no_rope_layers (`list[int]`, *optional*):
|
|
List with at least the same length as the number of layers in the model.
|
|
A `1` at an index position indicates that the corresponding layer will use RoPE,
|
|
while a `0` indicates that it's a NoPE layer.
|
|
no_rope_layer_interval (`int`, *optional*, defaults to 4):
|
|
If `no_rope_layers` is `None`, it will be created using a NoPE layer every
|
|
`no_rope_layer_interval` layers.
|
|
attention_chunk_size (`int`, *optional*, defaults to 8192):
|
|
Chunk size for the attention computation. Smaller value enforces more local attention and lowers memory.
|
|
layer_types (`list`, *optional*):
|
|
Attention pattern for each layer.
|
|
attn_temperature_tuning (`bool`, *optional*, defaults to `True`):
|
|
Whether to dynamically scale the attention temperature for each query token based on sequence length.
|
|
Recommended for long sequences (e.g., >32k tokens) to maintain stable output results.
|
|
floor_scale (`int`, *optional*, defaults to 8192):
|
|
Base scale (in tokens) for attention temperature tuning. Larger value delays scaling to longer positions.
|
|
attn_scale (`float`, *optional*, defaults to 0.1):
|
|
Strength of attention temperature tuning. Larger value increases scaling at long positions.
|
|
|
|
Example:
|
|
"""
|
|
|
|
model_type = "llama4_text"
|
|
keys_to_ignore_at_inference = ["past_key_values"]
|
|
default_theta = 500000.0
|
|
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.*.feed_forward.shared_expert.gate_proj": "colwise",
|
|
"layers.*.feed_forward.shared_expert.up_proj": "colwise",
|
|
"layers.*.feed_forward.shared_expert.down_proj": "rowwise",
|
|
"layers.*.feed_forward.experts.gate_up_proj": "packed_rowwise", # row because not linear
|
|
"layers.*.feed_forward.experts.down_proj": "colwise", # col because not linear
|
|
"layers.*.feed_forward.gate_proj": "colwise",
|
|
"layers.*.feed_forward.up_proj": "colwise",
|
|
"layers.*.feed_forward.down_proj": "rowwise",
|
|
}
|
|
base_model_ep_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.*.feed_forward.experts.gate_up_proj": "grouped_gemm", # row because not linear
|
|
"layers.*.feed_forward.experts.down_proj": "grouped_gemm", # col because not linear
|
|
"layers.*.feed_forward.gate_proj": "colwise",
|
|
"layers.*.feed_forward.up_proj": "colwise",
|
|
"layers.*.feed_forward.down_proj": "rowwise",
|
|
"layers.*.feed_forward.router": "ep_router",
|
|
}
|
|
|
|
def __init__(
|
|
self,
|
|
vocab_size=202048,
|
|
hidden_size=5120,
|
|
intermediate_size=8192,
|
|
intermediate_size_mlp=16384,
|
|
num_hidden_layers=48,
|
|
num_attention_heads=40,
|
|
num_key_value_heads=8,
|
|
head_dim=128,
|
|
hidden_act="silu",
|
|
max_position_embeddings=4096 * 32,
|
|
initializer_range=0.02,
|
|
rms_norm_eps=1e-5,
|
|
use_cache=True,
|
|
pad_token_id=None,
|
|
bos_token_id=1,
|
|
eos_token_id=2,
|
|
tie_word_embeddings=False,
|
|
attention_dropout=0.0,
|
|
num_experts_per_tok=1,
|
|
num_local_experts=16,
|
|
moe_layers=None,
|
|
interleave_moe_layer_step=1,
|
|
use_qk_norm=True,
|
|
output_router_logits=False,
|
|
router_aux_loss_coef=0.001,
|
|
router_jitter_noise=0.0,
|
|
rope_parameters: RopeParameters | dict[str, RopeParameters] | None = None,
|
|
no_rope_layers=None,
|
|
no_rope_layer_interval=4,
|
|
attention_chunk_size=8192,
|
|
layer_types=None,
|
|
attn_temperature_tuning=True,
|
|
floor_scale=8192,
|
|
attn_scale=0.1,
|
|
**kwargs,
|
|
):
|
|
self.tie_word_embeddings = tie_word_embeddings
|
|
self.pad_token_id = pad_token_id
|
|
self.bos_token_id = bos_token_id
|
|
self.eos_token_id = eos_token_id
|
|
self.attn_temperature_tuning = attn_temperature_tuning
|
|
self.attn_scale = attn_scale
|
|
self.floor_scale = floor_scale
|
|
self.vocab_size = vocab_size
|
|
self.max_position_embeddings = max_position_embeddings
|
|
self.hidden_size = hidden_size
|
|
self.intermediate_size = intermediate_size
|
|
self.intermediate_size_mlp = intermediate_size_mlp
|
|
self.num_hidden_layers = num_hidden_layers
|
|
self.num_attention_heads = num_attention_heads
|
|
self.attention_bias = False
|
|
# 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.attention_dropout = attention_dropout
|
|
self.head_dim = head_dim if head_dim is not None else self.hidden_size // self.num_attention_heads
|
|
self.use_qk_norm = use_qk_norm
|
|
self.num_experts_per_tok = num_experts_per_tok
|
|
self.num_local_experts = num_local_experts
|
|
|
|
self.output_router_logits = output_router_logits
|
|
self.router_aux_loss_coef = router_aux_loss_coef
|
|
self.router_jitter_noise = router_jitter_noise
|
|
|
|
# Backwards compatibility
|
|
if no_rope_layers == []:
|
|
no_rope_layers = None
|
|
|
|
default_no_rope_layers = [
|
|
int((layer_idx + 1) % no_rope_layer_interval != 0) for layer_idx in range(self.num_hidden_layers)
|
|
]
|
|
|
|
self.no_rope_layers = no_rope_layers if no_rope_layers else default_no_rope_layers
|
|
|
|
self.interleave_moe_layer_step = interleave_moe_layer_step
|
|
self.moe_layers = (
|
|
moe_layers
|
|
if moe_layers is not None
|
|
else list(
|
|
range(
|
|
interleave_moe_layer_step - 1,
|
|
num_hidden_layers,
|
|
interleave_moe_layer_step,
|
|
)
|
|
)
|
|
)
|
|
self.attention_chunk_size = attention_chunk_size
|
|
|
|
self.layer_types = layer_types
|
|
if layer_types is None:
|
|
self.layer_types = [
|
|
"chunked_attention" if no_rope else "full_attention" for no_rope in self.no_rope_layers
|
|
]
|
|
layer_type_validation(self.layer_types, self.num_hidden_layers)
|
|
|
|
self.rope_parameters = rope_parameters
|
|
super().__init__(**kwargs)
|
|
|
|
|
|
class Llama4Config(PreTrainedConfig):
|
|
r"""
|
|
This is the configuration class to store the configuration of a [`Llama4Model`]. It is used to instantiate an
|
|
Llama4 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 Llama4 109B.
|
|
|
|
e.g. [meta-llama/Llama-4-Scout-17B-16E](https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E)
|
|
|
|
Configuration objects inherit from [`PreTrainedConfig`] and can be used to control the model outputs. Read the
|
|
documentation from [`PreTrainedConfig`] for more information.
|
|
|
|
|
|
Args:
|
|
vision_config (`Llama4VisionConfig`, *optional*):
|
|
The Llama4 Vision config.
|
|
text_config (`Llama4TextConfig`, *optional*):
|
|
The Llama4 Text config.
|
|
boi_token_index (`int`, *optional*, defaults to 200080):
|
|
The begin-of-image token index to wrap the image prompt.
|
|
eoi_token_index (`int`, *optional*, defaults to 200081):
|
|
The end-of-image token index to wrap the image prompt.
|
|
image_token_index (`int`, *optional*, defaults to 200092):
|
|
The image token index to encode the image prompt.
|
|
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
|
Whether the model's input and output word embeddings should be tied.
|
|
|
|
```python
|
|
>>> from transformers import Llama4Model, Llama4Config
|
|
|
|
>>> # Initializing a Llama4 7B style configuration
|
|
>>> configuration = Llama4Config()
|
|
|
|
>>> # Initializing a model from the Llama4 7B style configuration
|
|
>>> model = Llama4Model(configuration)
|
|
|
|
>>> # Accessing the model configuration
|
|
>>> configuration = model.config
|
|
```"""
|
|
|
|
model_type = "llama4"
|
|
attribute_map = {
|
|
"image_token_id": "image_token_index",
|
|
"boi_token_id": "boi_token_index",
|
|
"eoi_token_id": "eoi_token_index",
|
|
}
|
|
sub_configs = {"text_config": Llama4TextConfig, "vision_config": Llama4VisionConfig}
|
|
base_model_tp_plan = {
|
|
"multi_modal_projector.linear_1": "colwise_rep",
|
|
}
|
|
|
|
def __init__(
|
|
self,
|
|
vision_config=None,
|
|
text_config=None,
|
|
boi_token_index=200080,
|
|
eoi_token_index=200081,
|
|
image_token_index=200092,
|
|
tie_word_embeddings=False,
|
|
**kwargs,
|
|
):
|
|
if vision_config is None:
|
|
self.vision_config = Llama4VisionConfig()
|
|
logger.info("vision_config is None, using default llama4 vision config")
|
|
elif isinstance(vision_config, dict):
|
|
self.vision_config = Llama4VisionConfig(**vision_config)
|
|
elif isinstance(vision_config, Llama4VisionConfig):
|
|
self.vision_config = vision_config
|
|
|
|
self.boi_token_index = boi_token_index
|
|
self.eoi_token_index = eoi_token_index
|
|
self.image_token_index = image_token_index
|
|
if text_config is None:
|
|
self.text_config = Llama4TextConfig()
|
|
logger.info("text_config is None, using default llama4 text config")
|
|
elif isinstance(text_config, dict):
|
|
self.text_config = Llama4TextConfig(**text_config)
|
|
elif isinstance(text_config, Llama4TextConfig):
|
|
self.text_config = text_config
|
|
|
|
self.tie_word_embeddings = tie_word_embeddings
|
|
super().__init__(**kwargs)
|
|
|
|
|
|
__all__ = ["Llama4Config", "Llama4TextConfig", "Llama4VisionConfig"]
|