ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

refactor glm edge (#12588)

Browse source
This commit is contained in:
Yishuo Wang 2024-12-20 15:36:57 +08:00 committed by GitHub
parent b0338c5529
commit 6ea8033635
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 8 additions and 44 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

52
python/llm/src/ipex_llm/transformers/models/glm.py
View file

@ -37,12 +37,12 @@ import torch
from typing import Optional, Tuple from typing import Optional, Tuple
from transformers.cache_utils import Cache from transformers.cache_utils import Cache
from transformers.models.glm.modeling_glm import repeat_kv, apply_rotary_pos_emb from transformers.models.glm.modeling_glm import apply_rotary_pos_emb
from ipex_llm.transformers.kv import DynamicNormalCache, DynamicFp8Cache from ipex_llm.transformers.kv import DynamicNormalCache, DynamicFp8Cache
from ipex_llm.transformers.models.common import merge_qkv_base from ipex_llm.transformers.models.common import merge_qkv_base
from ipex_llm.transformers.models.utils import use_sdp, use_sdp_causal from ipex_llm.transformers.models.common import scaled_dot_product_attention
from ipex_llm.transformers.models.utils import make_cache_contiguous_inplaced from ipex_llm.transformers.models.utils import make_cache_contiguous_inplaced
from ipex_llm.transformers.models.utils import use_quantize_kv_cache, restore_fp8_kv_cache from ipex_llm.transformers.models.utils import use_quantize_kv_cache
def merge_qkv(module: torch.nn.Module): def merge_qkv(module: torch.nn.Module):
@ -102,52 +102,16 @@ def glm_attention_forward(
else: else:
query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin) query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
use_quantizekv = isinstance(past_key_value, DynamicFp8Cache)
# sin and cos are specific to RoPE models; cache_position needed for the static cache # sin and cos are specific to RoPE models; cache_position needed for the static cache
cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position} cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
key_states, value_states = past_key_value.update(key_states, value_states, key_states, value_states = past_key_value.update(key_states, value_states,
self.layer_idx, cache_kwargs) self.layer_idx, cache_kwargs)
kv_seq_len = key_states.size(-2) attn_weights = None
if attention_mask is not None: # no matter the length, we just slice it attn_output = scaled_dot_product_attention(
attention_mask = attention_mask[:, :, :, : kv_seq_len] query_states, key_states, value_states,
attention_mask, q_len == key_states.size(2), self.scaling
if use_sdp(q_len, kv_seq_len, self.head_dim, query_states): )
import xe_addons
if use_quantizekv:
attn_output = xe_addons.sdp_fp8(query_states, key_states, value_states,
attention_mask)
else:
attn_output = xe_addons.sdp(query_states, key_states, value_states,
attention_mask)
elif use_sdp_causal(q_len, kv_seq_len, self.head_dim, query_states, self.training):
import xe_addons
if use_quantizekv:
attn_output = xe_addons.sdp_fp8_causal(query_states, key_states,
value_states, attention_mask)
else:
attn_output = xe_addons.sdp_causal(query_states, key_states,
value_states, attention_mask)
else:
if use_quantizekv:
key_states, value_states = restore_fp8_kv_cache(key_states, value_states,
query_states.dtype)
# repeat k/v heads if n_kv_heads < n_heads
key_states = repeat_kv(key_states, self.num_key_value_groups)
value_states = repeat_kv(value_states, self.num_key_value_groups)
attn_weights = torch.matmul(query_states,
key_states.transpose(2, 3)) * self.scaling
if attention_mask is not None:
attn_weights = attn_weights + attention_mask
# upcast attention to fp32
attn_weights = torch.nn.functional.softmax(attn_weights, dim=-1,
dtype=torch.float32).to(query_states.dtype)
attn_weights = torch.nn.functional.dropout(attn_weights, p=self.attention_dropout,
training=self.training)
attn_output = torch.matmul(attn_weights, value_states)
attn_output = attn_output.transpose(1, 2).contiguous() attn_output = attn_output.transpose(1, 2).contiguous()
attn_output = attn_output.reshape(bsz, q_len, self.hidden_size) attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)