[llama] add flash attn patch for npu (#5362)

applications/Colossal-LLaMA-2/colossal_llama2/utils/flash_attention_patch.py

@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
+import math
 from types import MethodType
 from typing import Optional, Tuple
 
 import torch
+import torch.nn as nn
 import torch.nn.functional as F
 from einops import rearrange
-from flash_attn.bert_padding import pad_input, unpad_input
+from transformers.models.llama.configuration_llama import LlamaConfig
-from flash_attn.flash_attn_interface import flash_attn_func, flash_attn_varlen_kvpacked_func
-from flash_attn.ops.rms_norm import rms_norm
 from transformers.models.llama.modeling_llama import (
     LlamaAttention,
     LlamaForCausalLM,
@@ -19,10 +19,15 @@ from transformers.models.llama.modeling_llama import (
     repeat_kv,
 )
 
+from colossalai.accelerator import get_accelerator
 from colossalai.logging import get_dist_logger
 
 logger = get_dist_logger()
 
+if get_accelerator().name == "cuda":
+    from flash_attn.bert_padding import pad_input, unpad_input
+    from flash_attn.flash_attn_interface import flash_attn_func, flash_attn_varlen_kvpacked_func
+    from flash_attn.ops.rms_norm import rms_norm
 
     def _prepare_decoder_attention_mask(
         self: LlamaModel,
@@ -51,7 +56,6 @@ def _prepare_decoder_attention_mask(
             return None  # Faster
         return attention_mask
 
-
     def attention_forward(
         self: LlamaAttention,
         hidden_states: torch.Tensor,
@@ -163,7 +167,9 @@ def attention_forward(
         if key_padding_mask is None:
             # (bsz, past_kv_len + q_len, num_heads, head_dim)
             output = flash_attn_func(q=q, k=k, v=v, dropout_p=0.0, softmax_scale=None, causal=True)  # (bsz, )
-        output = rearrange(output, pattern="... h d -> ... (h d)")  # (bsz, past_kv_len + q_len, num_heads * head_dim)
+            output = rearrange(
+                output, pattern="... h d -> ... (h d)"
+            )  # (bsz, past_kv_len + q_len, num_heads * head_dim)
         else:
             q, indices, cu_q_lens, max_q_len = unpad_input(hidden_states=q, attention_mask=key_padding_mask)
             kv, _, cu_kv_lens, max_kv_len = unpad_input(
@@ -194,14 +200,12 @@ def attention_forward(
         output = self.o_proj(output)  # (bsz, q_len, hidden_size)
         return output, None, past_key_value
 
-
     def rms_norm_forward(self: LlamaRMSNorm, hidden_states: torch.Tensor) -> torch.Tensor:
         """
         Formard function for RMS Norm
         """
         return rms_norm(x=hidden_states, weight=self.weight, epsilon=self.variance_epsilon)
 
-
     def replace_with_flash_attention(model: LlamaForCausalLM) -> None:
         for name, module in model.named_modules():
             if isinstance(module, LlamaAttention):
@@ -210,3 +214,139 @@ def replace_with_flash_attention(model: LlamaForCausalLM) -> None:
                 module._prepare_decoder_attention_mask = MethodType(_prepare_decoder_attention_mask, module)
             if isinstance(module, LlamaRMSNorm):
                 module.forward = MethodType(rms_norm_forward, module)
+
+elif get_accelerator().name == "npu":
+    import torch_npu
+
+    class NPULlamaAttention(LlamaAttention):
+        use_flash: bool = True
+
+        def __init__(self, config: LlamaConfig):
+            super().__init__(config)
+            self.setup()
+
+        def setup(self):
+            self._softmax_scale = 1 / math.sqrt(self.head_dim)
+
+        def forward(
+            self,
+            hidden_states: torch.Tensor,
+            attention_mask: Optional[torch.Tensor] = None,
+            position_ids: Optional[torch.LongTensor] = None,
+            past_key_value: Optional[Tuple[torch.Tensor]] = None,
+            output_attentions: bool = False,
+            use_cache: bool = False,
+        ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+            bsz, q_len, _ = hidden_states.size()
+
+            if self.config.pretraining_tp > 1:
+                key_value_slicing = (self.num_key_value_heads * self.head_dim) // self.config.pretraining_tp
+                query_slices = self.q_proj.weight.split(
+                    (self.num_heads * self.head_dim) // self.config.pretraining_tp, dim=0
+                )
+                key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
+                value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)
+
+                query_states = [F.linear(hidden_states, query_slices[i]) for i in range(self.config.pretraining_tp)]
+                query_states = torch.cat(query_states, dim=-1)
+
+                key_states = [F.linear(hidden_states, key_slices[i]) for i in range(self.config.pretraining_tp)]
+                key_states = torch.cat(key_states, dim=-1)
+
+                value_states = [F.linear(hidden_states, value_slices[i]) for i in range(self.config.pretraining_tp)]
+                value_states = torch.cat(value_states, dim=-1)
+
+            else:
+                query_states = self.q_proj(hidden_states)
+                key_states = self.k_proj(hidden_states)
+                value_states = self.v_proj(hidden_states)
+
+            query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+            key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+            value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+            kv_seq_len = key_states.shape[-2]
+            if past_key_value is not None:
+                kv_seq_len += past_key_value[0].shape[-2]
+            cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
+            query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
+
+            if past_key_value is not None:
+                # reuse k, v, self_attention
+                key_states = torch.cat([past_key_value[0], key_states], dim=2)
+                value_states = torch.cat([past_key_value[1], value_states], dim=2)
+
+            past_key_value = (key_states, value_states) if use_cache else None
+
+            key_states = repeat_kv(key_states, self.num_key_value_groups)
+            value_states = repeat_kv(value_states, self.num_key_value_groups)
+
+            if not self.use_flash:
+                attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
+
+                if attn_weights.size() != (bsz, self.num_heads, q_len, kv_seq_len):
+                    raise ValueError(
+                        f"Attention weights should be of size {(bsz, self.num_heads, q_len, kv_seq_len)}, but is"
+                        f" {attn_weights.size()}"
+                    )
+
+                if attention_mask is not None:
+                    if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
+                        raise ValueError(
+                            f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
+                        )
+                    attn_weights = attn_weights + attention_mask
+
+                # upcast attention to fp32
+                attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
+                attn_output = torch.matmul(attn_weights, value_states)
+            else:
+                attn_output, *_ = torch_npu.npu_fusion_attention(
+                    query_states,
+                    key_states,
+                    value_states,
+                    self.num_heads,
+                    "BNSD",
+                    atten_mask=attention_mask.bool(),
+                    scale=self._softmax_scale,
+                    padding_mask=None,
+                    pre_tockens=65535,
+                    next_tockens=0,
+                    keep_prob=1.0,
+                    inner_precise=0,
+                )
+
+            if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
+                raise ValueError(
+                    f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
+                    f" {attn_output.size()}"
+                )
+
+            attn_output = attn_output.transpose(1, 2).contiguous()
+            attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
+
+            if self.config.pretraining_tp > 1:
+                attn_output = attn_output.split(self.hidden_size // self.config.pretraining_tp, dim=2)
+                o_proj_slices = self.o_proj.weight.split(self.hidden_size // self.config.pretraining_tp, dim=1)
+                attn_output = sum(
+                    [F.linear(attn_output[i], o_proj_slices[i]) for i in range(self.config.pretraining_tp)]
+                )
+            else:
+                attn_output = self.o_proj(attn_output)
+
+            if not output_attentions:
+                attn_weights = None
+
+            return attn_output, attn_weights, past_key_value
+
+    class NPURMSNorm(LlamaRMSNorm):
+        def forward(self, hidden_states):
+            return torch_npu.npu_rms_norm(hidden_states, self.weight, epsilon=self.variance_epsilon)[0]
+
+    def replace_with_flash_attention(model: LlamaForCausalLM) -> None:
+        for name, module in model.named_modules():
+            if isinstance(module, LlamaAttention):
+                module.__class__ = NPULlamaAttention
+                module.setup()
+            if isinstance(module, LlamaRMSNorm):
+                module.__class__ = NPURMSNorm