# Copyright 2023 The EASYDEL Author @erfanzar (Erfan Zare Chavoshi).
#
# 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
#
# https://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.
import typing as tp
from jax.sharding import PartitionSpec
from easydel.infra.base_module import EasyDeLBaseConfig
from easydel.infra.etils import EasyDeLGradientCheckPointers
from easydel.infra.factory import register_config
[docs]@register_config("phimoe")
class PhiMoeConfig(EasyDeLBaseConfig):
"""
Configuration objects inherit from [`EasyDeLBaseConfig`] and can be used to control the model outputs. Read
the documentation from [`EasyDeLBaseConfig`] for more information.
Args:
vocab_size (`int`, *optional*, defaults to 32064):
Vocabulary size of the PhiMoE model. Defines the number of different tokens that can be represented by the
`inputs_ids` passed when calling [`PhiMoEModel`]
hidden_size (`int`, *optional*, defaults to 4096):
Dimension of the hidden representations.
intermediate_size (`int`, *optional*, defaults to 6400):
Dimension of the MLP representations.
num_hidden_layers (`int`, *optional*, defaults to 32):
Number of hidden layers in the Transformer encoder.
num_attention_heads (`int`, *optional*, defaults to 32):
Number of attention heads for each attention layer in the Transformer encoder.
num_key_value_heads (`int`, *optional*, defaults to 8):
This is the number of key_value heads that should be used to implement Grouped Query Attention. If
`num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
`num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
by meanpooling all the original heads within that group. For more details checkout [this
paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to `8`.
hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
The non-linear activation function (function or string) in the decoder.
max_position_embeddings (`int`, *optional*, defaults to `4096*32`):
The maximum sequence length that this model might ever be used with. Mixtral's sliding window attention
allows sequence of up to 4096*32 tokens.
initializer_range (`float`, *optional*, defaults to 0.02):
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
rms_norm_eps (`float`, *optional*, defaults to 1e-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 (not used by all models). Only
relevant if `config.is_decoder=True`.
pad_token_id (`int`, *optional*):
The id of the padding token.
bos_token_id (`int`, *optional*, defaults to 1):
The id of the "beginning-of-sequence" token.
eos_token_id (`int`, *optional*, defaults to 2):
The id of the "end-of-sequence" token.
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
Whether the model's input and output word embeddings should be tied.
rope_theta (`float`, *optional*, defaults to 10000.0):
The base period of the RoPE embeddings.
rope_scaling (`dict`, *optional*):
The scaling strategy for the RoPE embeddings. If `None`, no scaling is applied. If a dictionary, it must
contain the following keys: `type`, `short_factor`, `long_factor`, `short_mscale`, `long_mscale` and
`original_max_position_embeddings`. The `type` must be `longrope`, the `short_mscale` and `long_scale` must
be numbers, the `short_factor` and `long_factor` must be lists of numbers with the same length as half of
the attention head size and the `original_max_position_embeddings` must be an integer.
sliding_window (`int`, *optional*):
Sliding window attention window size. If not specified, will default to `262144`.
attention_dropout (`float`, *optional*, defaults to 0.0):
The dropout ratio for the attention probabilities.
num_experts_per_tok (`int`, *optional*, defaults to 2):
The number of experts to root per-token, can be also interpreted as the `top-p` routing
parameter
num_local_experts (`int`, *optional*, defaults to 16):
Number of experts per Sparse MLP layer.
output_router_logits (`bool`, *optional*, defaults to `False`):
Whether or not the router logits should be returned by the model. Enabeling this will also
allow the model to output the auxiliary loss. See [here]() for more details
router_aux_loss_coef (`float`, *optional*, defaults to 0.0):
The aux loss factor for the total loss.
router_jitter_noise (`float`, *optional*, defaults to 0.01):
Amount of noise to add to the router.
bits (`int`, *optional*):
The number of bits to quantize the model to.
gradient_checkpointing (`str`, *optional*, defaults to `"nothing_saveable"`):
The gradient checkpointing configuration.
"""
model_type: str = "phimoe"
def __init__(
self,
vocab_size=32064,
hidden_size=4096,
intermediate_size=6400,
num_hidden_layers=32,
num_attention_heads=32,
num_key_value_heads=8,
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,
rope_theta=1e6,
rope_scaling=None,
sliding_window=None,
attention_dropout=0.0,
num_experts_per_tok=2,
num_local_experts=16,
output_router_logits=False,
router_aux_loss_coef=0.001,
router_jitter_noise=0.01,
input_jitter_noise=0.0,
attention_bias=False,
embd_pdrop: float = 0.0,
lm_head_bias=False,
bits: tp.Optional[int] = None,
gradient_checkpointing: EasyDeLGradientCheckPointers = EasyDeLGradientCheckPointers.NONE,
**kwargs,
) -> None:
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
self.sliding_window = sliding_window
self.attention_bias = attention_bias
self.lm_head_bias = lm_head_bias
# 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_dropout = attention_dropout
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
self.input_jitter_noise = input_jitter_noise
self.embd_pdrop = embd_pdrop
self.rope_scaling = rope_scaling or {}
self._rope_scaling_validation()
self.bits = bits
self.gradient_checkpointing = gradient_checkpointing
super().__init__(
pad_token_id=pad_token_id,
bos_token_id=bos_token_id,
eos_token_id=eos_token_id,
tie_word_embeddings=tie_word_embeddings,
bits=bits,
**kwargs,
)
[docs] def attach_custom_arguments(
self,
bits: tp.Optional[int] = None,
embd_pdrop: float = 0.0,
gradient_checkpointing: EasyDeLGradientCheckPointers = EasyDeLGradientCheckPointers.NONE,
**kwargs,
):
self.bits = bits
self.embd_pdrop = embd_pdrop
self.gradient_checkpointing = gradient_checkpointing
for k, v in kwargs.items():
if not hasattr(self, k):
setattr(self, k, v)
[docs] def get_partition_rules(self, fully_sharded_data_parallel: bool = True):
"""
Get the partition rules for the model.
Args:
fully_sharded_data_parallel (`bool`, *optional*, defaults to `True`):
Whether to use fully sharded data parallelism.
Returns:
`tp.Tuple[tp.Tuple[str, PartitionSpec]]`: The partition rules.
"""
return (
(
("embed_tokens/embedding", PartitionSpec(("fsdp", "sp"), "tp")),
(
"norm/kernel",
PartitionSpec(
("fsdp", "sp"),
),
),
(
"post_attention_layernorm/kernel",
PartitionSpec(
("fsdp", "sp"),
),
),
(
"input_layernorm/kernel",
PartitionSpec(
("fsdp", "sp"),
),
),
("mlp/gate_up_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("mlp/down_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("self_attn/o_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("self_attn/qkv_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("lm_head/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
(
".*",
PartitionSpec(
None,
),
),
)
if fully_sharded_data_parallel
else (
("embed_tokens/embedding", PartitionSpec(("fsdp", "sp"), "tp")),
(
"norm/kernel",
PartitionSpec(
None,
),
),
(
"post_attention_layernorm/kernel",
PartitionSpec(
None,
),
),
(
"input_layernorm/kernel",
PartitionSpec(
None,
),
),
("mlp/gate_up_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("mlp/down_proj/kernel", PartitionSpec("tp", ("fsdp", "sp"))),
(
"self_attn/o_proj/kernel",
PartitionSpec(
"tp",
("fsdp", "sp"),
),
),
("self_attn/qkv_proj/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
("lm_head/kernel", PartitionSpec(("fsdp", "sp"), "tp")),
(
".*",
PartitionSpec(
None,
),
),
)
)
def _rope_scaling_validation(self):
"""
Validate the `rope_scaling` configuration.
"""
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 6:
raise ValueError(
"`rope_scaling` must be a dictionary with three fields, `type`, `short_factor`, `long_factor`, "
f"`short_mscale`, `long_mscale` and `original_max_position_embeddings`, got {self.rope_scaling}"
)
rope_scaling_type = self.rope_scaling.get("type", None)
rope_scaling_short_factor = self.rope_scaling.get("short_factor", None)
rope_scaling_long_factor = self.rope_scaling.get("long_factor", None)
rope_scaling_short_mscale = self.rope_scaling.get("short_mscale", None)
rope_scaling_long_mscale = self.rope_scaling.get("long_mscale", None)
original_max_position_embeddings = self.rope_scaling.get(
"original_max_position_embeddings", None
)
if rope_scaling_type is None or rope_scaling_type not in ["longrope"]:
raise ValueError(
f"`rope_scaling`'s type field must be one of ['longrope'], got {rope_scaling_type}"
)
if not (
isinstance(rope_scaling_short_factor, list)
and all(isinstance(x, (int, float)) for x in rope_scaling_short_factor)
):
raise ValueError(
f"`rope_scaling`'s short_factor field must be a list of numbers, got {rope_scaling_short_factor}"
)
if (
not len(rope_scaling_short_factor)
== self.hidden_size // self.num_attention_heads // 2
):
raise ValueError(
f"`rope_scaling`'s short_factor field must have length {self.hidden_size // self.num_attention_heads // 2}, got {len(rope_scaling_short_factor)}"
)
if not (
isinstance(rope_scaling_long_factor, list)
and all(isinstance(x, (int, float)) for x in rope_scaling_long_factor)
):
raise ValueError(
f"`rope_scaling`'s long_factor field must be a list of numbers, got {rope_scaling_long_factor}"
)
if (
not len(rope_scaling_long_factor)
== self.hidden_size // self.num_attention_heads // 2
):
raise ValueError(
f"`rope_scaling`'s long_factor field must have length {self.hidden_size // self.num_attention_heads // 2}, got {len(rope_scaling_long_factor)}"
)
if not isinstance(rope_scaling_short_mscale, (int, float)):
raise ValueError(
f"`rope_scaling`'s short_mscale field must be a number, got {rope_scaling_short_mscale}"
)
if not isinstance(rope_scaling_long_mscale, (int, float)):
raise ValueError(
f"`rope_scaling`'s long_mscale field must be a number, got {rope_scaling_long_mscale}"
)
if not isinstance(original_max_position_embeddings, int):
raise ValueError(
f"`rope_scaling`'s original_max_position_embeddings field must be an integer, got {original_max_position_embeddings}"
)
@property
def granted_freq_max_position_embedding(self) -> int:
return getattr(
self,
"freq_max_position_embeddings",
self.max_position_embeddings,
)
@property
def granted_mask_max_position_embedding(self) -> int:
return getattr(
self,
"mask_max_position_embeddings",
self.max_position_embeddings,
)