mirror of
https://github.com/ml-explore/mlx-examples.git
synced 2025-06-24 09:21:18 +08:00
add internlm3 (#1206)
This commit is contained in:
Awni Hannun
GitHub
parent
6ae6c72c2e
commit
50f0a7f6d9
241
llms/mlx_lm/models/internlm3.py
Normal file
241
llms/mlx_lm/models/internlm3.py
Normal file
@ -0,0 +1,241 @@
|
||||
# Copyright © 2023-2024 Apple Inc.
|
||||
|
||||
from dataclasses import dataclass
|
||||
from typing import Any, Dict, Optional, Tuple, Union
|
||||
|
||||
import mlx.core as mx
|
||||
import mlx.nn as nn
|
||||
|
||||
from .base import BaseModelArgs, create_attention_mask, scaled_dot_product_attention
|
||||
|
||||
|
||||
@dataclass
|
||||
class ModelArgs(BaseModelArgs):
|
||||
model_type: str
|
||||
hidden_size: int
|
||||
num_hidden_layers: int
|
||||
intermediate_size: int
|
||||
num_attention_heads: int
|
||||
rms_norm_eps: float
|
||||
vocab_size: int
|
||||
bias: bool = False
|
||||
qkv_bias: bool = False
|
||||
max_position_embeddings: int = 32768
|
||||
num_key_value_heads: int = None
|
||||
rope_theta: float = 10000
|
||||
rope_traditional: bool = False
|
||||
rope_scaling: Optional[Dict[str, Union[float, str]]] = None
|
||||
tie_word_embeddings: bool = False
|
||||
|
||||
def __post_init__(self):
|
||||
if self.num_key_value_heads is None:
|
||||
self.num_key_value_heads = self.num_attention_heads
|
||||
|
||||
if self.rope_scaling:
|
||||
required_keys = {"factor", "rope_type"}
|
||||
if not all(key in self.rope_scaling for key in required_keys):
|
||||
raise ValueError(f"rope_scaling must contain keys {required_keys}")
|
||||
|
||||
if self.rope_scaling["rope_type"] not in ["linear", "dynamic"]:
|
||||
raise ValueError(
|
||||
"rope_scaling 'rope_type' currently only supports 'linear' or 'dynamic"
|
||||
)
|
||||
|
||||
|
||||
class DynamicNTKScalingRoPE(nn.Module):
|
||||
"""Implements the rotary positional encoding with Dynamic NTK scaling."""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
dims: int,
|
||||
max_position_embeddings: int = 2048,
|
||||
traditional: bool = False,
|
||||
base: float = 10000,
|
||||
scale: float = 1.0,
|
||||
):
|
||||
super().__init__()
|
||||
self.max_position_embeddings = max_position_embeddings
|
||||
self.original_base = base
|
||||
self.dims = dims
|
||||
self.traditional = traditional
|
||||
self.scale = scale
|
||||
|
||||
def extra_repr(self):
|
||||
return f"{self.dims}, traditional={self.traditional}, max_position_embeddings={self.max_position_embeddings}, scaling_factor={self.scaling_factor}"
|
||||
|
||||
def __call__(self, x, offset: int = 0):
|
||||
seq_len = x.shape[1] + offset
|
||||
if seq_len > self.max_position_embeddings:
|
||||
base = self.original_base * (
|
||||
(self.scale * seq_len / self.max_position_embeddings) - (self.scale - 1)
|
||||
) ** (self.dims / (self.dims - 2))
|
||||
else:
|
||||
base = self.original_base
|
||||
|
||||
return mx.fast.rope(
|
||||
x,
|
||||
self.dims,
|
||||
traditional=self.traditional,
|
||||
base=base,
|
||||
scale=self.scale,
|
||||
offset=offset,
|
||||
)
|
||||
|
||||
|
||||
class Attention(nn.Module):
|
||||
def __init__(self, args: ModelArgs):
|
||||
super().__init__()
|
||||
|
||||
dim = args.hidden_size
|
||||
qkv_bias = args.qkv_bias
|
||||
self.n_heads = n_heads = args.num_attention_heads
|
||||
self.n_kv_heads = n_kv_heads = args.num_key_value_heads
|
||||
self.n_kv_groups = n_heads // args.num_key_value_heads
|
||||
|
||||
self.head_dim = head_dim = args.hidden_size // n_heads
|
||||
self.scale = head_dim**-0.5
|
||||
|
||||
self.q_proj = nn.Linear(dim, n_heads * head_dim, bias=qkv_bias)
|
||||
self.k_proj = nn.Linear(dim, n_kv_heads * head_dim, bias=qkv_bias)
|
||||
self.v_proj = nn.Linear(dim, n_kv_heads * head_dim, bias=qkv_bias)
|
||||
self.o_proj = nn.Linear(n_heads * head_dim, dim, bias=qkv_bias)
|
||||
|
||||
rope_scale = (
|
||||
1 / args.rope_scaling["factor"]
|
||||
if args.rope_scaling is not None
|
||||
and args.rope_scaling["rope_type"] == "linear"
|
||||
else 2.0
|
||||
)
|
||||
|
||||
self.rope = DynamicNTKScalingRoPE(
|
||||
head_dim,
|
||||
max_position_embeddings=args.max_position_embeddings,
|
||||
traditional=args.rope_traditional,
|
||||
base=args.rope_theta,
|
||||
scale=rope_scale,
|
||||
)
|
||||
|
||||
def __call__(
|
||||
self,
|
||||
x: mx.array,
|
||||
mask: Optional[mx.array] = None,
|
||||
cache: Optional[Any] = None,
|
||||
) -> mx.array:
|
||||
B, L, D = x.shape
|
||||
|
||||
queries, keys, values = self.q_proj(x), self.k_proj(x), self.v_proj(x)
|
||||
|
||||
# Prepare the queries, keys and values for the attention computation
|
||||
queries = queries.reshape(B, L, self.n_heads, -1).transpose(0, 2, 1, 3)
|
||||
keys = keys.reshape(B, L, self.n_kv_heads, -1).transpose(0, 2, 1, 3)
|
||||
values = values.reshape(B, L, self.n_kv_heads, -1).transpose(0, 2, 1, 3)
|
||||
|
||||
if cache is not None:
|
||||
queries = self.rope(queries, offset=cache.offset)
|
||||
keys = self.rope(keys, offset=cache.offset)
|
||||
keys, values = cache.update_and_fetch(keys, values)
|
||||
else:
|
||||
queries = self.rope(queries)
|
||||
keys = self.rope(keys)
|
||||
|
||||
output = scaled_dot_product_attention(
|
||||
queries, keys, values, cache=cache, scale=self.scale, mask=mask
|
||||
)
|
||||
|
||||
output = output.transpose(0, 2, 1, 3).reshape(B, L, -1)
|
||||
return self.o_proj(output)
|
||||
|
||||
|
||||
class MLP(nn.Module):
|
||||
def __init__(self, dim, hidden_dim, bias):
|
||||
super().__init__()
|
||||
self.gate_proj = nn.Linear(dim, hidden_dim, bias=bias)
|
||||
self.down_proj = nn.Linear(hidden_dim, dim, bias=bias)
|
||||
self.up_proj = nn.Linear(dim, hidden_dim, bias=bias)
|
||||
|
||||
def __call__(self, x) -> mx.array:
|
||||
return self.down_proj(nn.silu(self.gate_proj(x)) * self.up_proj(x))
|
||||
|
||||
|
||||
class TransformerBlock(nn.Module):
|
||||
def __init__(self, args: ModelArgs):
|
||||
super().__init__()
|
||||
self.self_attn = Attention(args)
|
||||
self.mlp = MLP(args.hidden_size, args.intermediate_size, args.bias)
|
||||
self.input_layernorm = nn.RMSNorm(args.hidden_size, eps=args.rms_norm_eps)
|
||||
self.post_attention_layernorm = nn.RMSNorm(
|
||||
args.hidden_size, eps=args.rms_norm_eps
|
||||
)
|
||||
|
||||
def __call__(
|
||||
self,
|
||||
x: mx.array,
|
||||
mask: Optional[mx.array] = None,
|
||||
cache: Optional[Any] = None,
|
||||
) -> mx.array:
|
||||
r = self.self_attn(self.input_layernorm(x), mask, cache)
|
||||
h = x + r
|
||||
r = self.mlp(self.post_attention_layernorm(h))
|
||||
out = h + r
|
||||
return out
|
||||
|
||||
|
||||
class InternLM2Model(nn.Module):
|
||||
def __init__(self, args: ModelArgs):
|
||||
super().__init__()
|
||||
assert args.vocab_size > 0
|
||||
self.embed_tokens = nn.Embedding(args.vocab_size, args.hidden_size)
|
||||
self.layers = [
|
||||
TransformerBlock(args=args) for _ in range(args.num_hidden_layers)
|
||||
]
|
||||
self.norm = nn.RMSNorm(args.hidden_size, eps=args.rms_norm_eps)
|
||||
|
||||
def __call__(
|
||||
self,
|
||||
inputs: mx.array,
|
||||
mask: mx.array = None,
|
||||
cache=None,
|
||||
):
|
||||
h = self.embed_tokens(inputs)
|
||||
|
||||
if mask is None:
|
||||
mask = create_attention_mask(h, cache)
|
||||
|
||||
if cache is None:
|
||||
cache = [None] * len(self.layers)
|
||||
|
||||
for layer, c in zip(self.layers, cache):
|
||||
h = layer(h, mask, cache=c)
|
||||
|
||||
return self.norm(h)
|
||||
|
||||
|
||||
class Model(nn.Module):
|
||||
def __init__(self, args: ModelArgs):
|
||||
super().__init__()
|
||||
self.args = args
|
||||
self.model_type = args.model_type
|
||||
self.model = InternLM2Model(args)
|
||||
if not args.tie_word_embeddings:
|
||||
self.lm_head = nn.Linear(args.hidden_size, args.vocab_size, bias=False)
|
||||
|
||||
def __call__(
|
||||
self,
|
||||
inputs: mx.array,
|
||||
mask: mx.array = None,
|
||||
cache=None,
|
||||
):
|
||||
out = self.model(inputs, mask, cache)
|
||||
if self.args.tie_word_embeddings:
|
||||
out = self.model.embed_tokens.as_linear(out)
|
||||
else:
|
||||
out = self.lm_head(out)
|
||||
return out
|
||||
|
||||
def sanitize(self, weights):
|
||||
# Remove unused precomputed rotary freqs
|
||||
return {k: v for k, v in weights.items() if "attention.rope.inv_freq" not in k}
|
||||
|
||||
@property
|
||||
def layers(self):
|
||||
return self.model.layers
|
||||
@ -100,6 +100,7 @@ def linear_to_lora_layers(
|
||||
"minicpm",
|
||||
"deepseek",
|
||||
"olmo2",
|
||||
"internlm3",
|
||||
]:
|
||||
keys = set(["self_attn.q_proj", "self_attn.v_proj"])
|
||||
if model.model_type in ["mixtral", "phimoe"]:
|
||||
|
||||
@ -927,6 +927,23 @@ class TestModels(unittest.TestCase):
|
||||
model, args.model_type, args.vocab_size, args.num_hidden_layers
|
||||
)
|
||||
|
||||
def test_internlm3(self):
|
||||
from mlx_lm.models import internlm3
|
||||
|
||||
args = internlm3.ModelArgs(
|
||||
model_type="internlm3",
|
||||
hidden_size=1024,
|
||||
num_hidden_layers=4,
|
||||
intermediate_size=2048,
|
||||
num_attention_heads=4,
|
||||
rms_norm_eps=1e-5,
|
||||
vocab_size=10_000,
|
||||
)
|
||||
model = internlm3.Model(args)
|
||||
self.model_test_runner(
|
||||
model, args.model_type, args.vocab_size, args.num_hidden_layers
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
|
||||
Loading…
Reference in New Issue
Block a user