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

use new rotary two in chatglm4 (#11312)

Browse source 
* use new rotary two in chatglm4

* rempve
This commit is contained in:
Xin Qiu 2024-06-13 19:02:18 +08:00 committed by GitHub
parent f1410d6823
commit 1b0c4c8cb8
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 32 additions and 65 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

97
python/llm/src/ipex_llm/transformers/models/chatglm4.py
View file

@ -47,10 +47,6 @@ def chatglm4_model_forward(
) -> Union[Tuple, BaseModelOutputWithPast]: ) -> Union[Tuple, BaseModelOutputWithPast]:
from ipex_llm.transformers.kv import DynamicFp8Cache from ipex_llm.transformers.kv import DynamicFp8Cache
use_cache = use_cache if use_cache is not None else self.config.use_cache use_cache = use_cache if use_cache is not None else self.config.use_cache
# if use_cache and use_quantize_kv_cache(
# self.encoder.layers[0].self_attention.query_key_value, input_ids):
# if not isinstance(past_key_values, DynamicFp8Cache):
# past_key_values = DynamicFp8Cache.from_legacy_cache(past_key_values)
return chatglm4_model_forward_internal( return chatglm4_model_forward_internal(
self=self, self=self,
input_ids=input_ids, input_ids=input_ids,
@ -108,25 +104,17 @@ def chatglm4_model_forward_internal(
dtype=torch.int64, device=inputs_embeds.device) dtype=torch.int64, device=inputs_embeds.device)
position_ids = position_ids.repeat(batch_size, 1) position_ids = position_ids.repeat(batch_size, 1)
use_fuse_rope = input_ids.device.type == "xpu" if getattr(self.rotary_pos_emb, "cached_dtype", None) != inputs_embeds.dtype:
use_fuse_rope = use_fuse_rope and not self.training rot_dim = self.rotary_pos_emb.dim
base = 10000 * getattr(self.rotary_pos_emb, "rope_ratio", 1)
# Rotary positional embeddings # We should generate float inv_freq to avoid overflow, as base is too large.
rotary_pos_emb = self.rotary_pos_emb(self.seq_length) inv_freq = 1.0 / (base ** (torch.arange(0, rot_dim, 2,
if position_ids is not None: dtype=torch.float,
rotary_pos_emb = rotary_pos_emb[position_ids] device=inputs_embeds.device) / rot_dim))
else: self.rotary_pos_emb.register_buffer("inv_freq",
rotary_pos_emb = rotary_pos_emb[None, :seq_length] inv_freq.to(inputs_embeds.dtype),
if use_fuse_rope: persistent=False)
# Repeat cos sin here, call only once for each token. self.rotary_pos_emb.cached = True
# Chatglm2's rotary embedding is similar to gptj's, is rotate_every_two.
# If put this to attension forward, it will generate too many times.
cos, sin = rotary_pos_emb.split(rotary_pos_emb.shape[-1] // 2, dim=-1)
cos = cos.squeeze(-1)
sin = sin.squeeze(-1)
cos = torch.repeat_interleave(cos[:, :, None, :], 2, 3)
sin = torch.repeat_interleave(sin[:, :, None, :], 2, 3)
rotary_pos_emb = (cos, sin)
# `full_attention_mask` is not None only when # `full_attention_mask` is not None only when
# `past_key_values` is not None and `seq_length` > 1 # `past_key_values` is not None and `seq_length` > 1
@ -148,7 +136,7 @@ def chatglm4_model_forward_internal(
hidden_states, presents, all_hidden_states, all_self_attentions = self.encoder( hidden_states, presents, all_hidden_states, all_self_attentions = self.encoder(
inputs_embeds, causal_mask, inputs_embeds, causal_mask,
rotary_pos_emb=rotary_pos_emb, rotary_pos_emb=(self.rotary_pos_emb.inv_freq, position_ids),
kv_caches=past_key_values, use_cache=use_cache, output_hidden_states=output_hidden_states kv_caches=past_key_values, use_cache=use_cache, output_hidden_states=output_hidden_states
) )
# ipex-llm changes end # ipex-llm changes end
@ -172,26 +160,6 @@ def chatglm4_model_forward_internal(
) )
def apply_rotary_pos_emb(x: torch.Tensor, rope_cache: torch.Tensor) -> torch.Tensor:
# x: [b, np, sq, hn]
b, np, sq, hn = x.size(0), x.size(1), x.size(2), x.size(3)
rot_dim = rope_cache.shape[-2] * 2
x, x_pass = x[..., :rot_dim], x[..., rot_dim:]
# truncate to support variable sizes
rope_cache = rope_cache[:, :sq]
xshaped = x.reshape(b, np, sq, rot_dim // 2, 2)
rope_cache = rope_cache.view(-1, 1, sq, xshaped.size(3), 2)
x_out2 = torch.stack(
[
xshaped[..., 0] * rope_cache[..., 0] - xshaped[..., 1] * rope_cache[..., 1],
xshaped[..., 1] * rope_cache[..., 0] + xshaped[..., 0] * rope_cache[..., 1],
],
-1,
)
x_out2 = x_out2.flatten(3)
return torch.cat((x_out2, x_pass), dim=-1)
def chatglm4_attention_forward( def chatglm4_attention_forward(
self, hidden_states, attention_mask, rotary_pos_emb, kv_cache=None, use_cache=True self, hidden_states, attention_mask, rotary_pos_emb, kv_cache=None, use_cache=True
): ):
@ -209,34 +177,33 @@ def chatglm4_attention_forward(
qkv = self.query_key_value(hidden_states) qkv = self.query_key_value(hidden_states)
# [bs, q_len, np * 3 * hn] -> [bsz, n_head, seq_len, head_dim] # [bs, q_len, np * 3 * hn] -> [bsz, n_head, seq_len, head_dim]
qkv = qkv.view(bsz, q_len, n_head + 2 * n_kv_head, head_dim) qkv = qkv.view(bsz, q_len, n_head + 2 * n_kv_head, head_dim)
qkv = qkv.transpose(1, 2)
query_states, key_states, value_states = qkv.split([n_head, query_states, key_states, value_states = qkv.split([n_head,
n_kv_head, n_kv_head,
n_kv_head], dim=2) n_kv_head], dim=1)
kv_seq_len = key_states.shape[1] kv_seq_len = key_states.shape[2]
if past_key_value is not None: if past_key_value is not None:
kv_seq_len += past_key_value[0].shape[2] kv_seq_len += past_key_value[0].shape[2]
if isinstance(rotary_pos_emb, tuple) and len(rotary_pos_emb) == 2: # IPEX-LLM OPT: fuse rope
# use_fuse_rope, see chatglm4_model_forward inv_freq, position_ids = rotary_pos_emb
cos, sin = rotary_pos_emb rot_dim = inv_freq.size(-1) * 2
rot_dim = cos.shape[-1] if should_use_fuse_rope(hidden_states, rotary_pos_emb[1], self.training):
query_layer_cur = query_states[..., :rot_dim] import xe_addons
key_layer_cur = key_states[..., :rot_dim] xe_addons.rotary_two_inplaced(inv_freq, position_ids,
# ipex_llm's apply_rotary_embedding can change the origin storage, query_states[..., :rot_dim], key_states[..., :rot_dim])
# so query_layer will get the result directly. else:
torch.ops.torch_ipex.apply_rotary_embedding(query_layer_cur, sin, cos, query_layer_cur) idx_theta = torch.outer(position_ids[0].float(),
torch.ops.torch_ipex.apply_rotary_embedding(key_layer_cur, sin, cos, key_layer_cur) inv_freq.float()).to(hidden_states.dtype)
query_states = query_states.transpose(1, 2) idx_theta = idx_theta.unsqueeze(0).unsqueeze(0)
key_states = key_states.transpose(1, 2) cos = torch.cos(idx_theta).repeat_interleave(2, -1)
value_states = value_states.transpose(1, 2) sin = torch.sin(idx_theta).repeat_interleave(2, -1)
elif rotary_pos_emb is not None: q_rot, k_rot = apply_rotary_pos_emb(query_states[..., :rot_dim], key_states[..., :rot_dim],
query_states = query_states.transpose(1, 2) cos, sin, position_ids, "chatglm")
key_states = key_states.transpose(1, 2) query_states[..., :rot_dim] = q_rot[...]
value_states = value_states.transpose(1, 2) key_states[..., :rot_dim] = k_rot[...]
query_states = apply_rotary_pos_emb(query_states, rotary_pos_emb)
key_states = apply_rotary_pos_emb(key_states, rotary_pos_emb)
# IPEX-LLM OPT: kv cache and quantize kv # IPEX-LLM OPT: kv cache and quantize kv
use_quantize_kv = use_quantize_kv_cache(self.query_key_value, hidden_states) use_quantize_kv = use_quantize_kv_cache(self.query_key_value, hidden_states)