add basic support for Baichuan-M1-14B-Instruct (#12808)

python/llm/src/ipex_llm/transformers/convert.py

@@ -1062,6 +1062,11 @@ def _optimize_pre(model, qtype=None):
         from ipex_llm.transformers.models.glm import merge_qkv, split_mlp
         model.apply(merge_qkv)
         model.apply(split_mlp)
+    elif model.config.model_type == "baichuan_m1":
+        from ipex_llm.transformers.models.baichuan_m1 import pre_register_inv_freq
+        model.apply(pre_register_inv_freq)
+    elif model.config.model_type == "multi_modality":
+        pass
 
     return model
 
@@ -1994,5 +1999,21 @@ def _optimize_post(model):
         model.llm.config.rope_scaling = {"rope_type": "default"}
         _optimize_post(model.llm)
         model.llm.config.model_type = "megrezo"
+    elif model.config.model_type == "baichuan_m1":
+        modeling_module_name = model.__class__.__module__
+        module = importlib.import_module(modeling_module_name)
+        from ipex_llm.transformers.models.common import rms_norm_forward
+        from ipex_llm.transformers.models.baichuan_m1 import model_forward
+        from ipex_llm.transformers.models.baichuan_m1 import eager_attention_forward
+        convert_forward(model, module.BaichuanModel, model_forward)
+        convert_forward(model, module.BaichuanRMSNorm, rms_norm_forward)
+        convert_forward(model, module.BaichuanAttention, eager_attention_forward)
+    elif model.config.model_type == "multi_modality":
+        # vision
+        vpm_modeling_module_name = model.vision_model.vision_tower.__class__.__module__
+        vpm_module = importlib.import_module(vpm_modeling_module_name)
+
+        from ipex_llm.transformers.models.janus import vision_attention_forward
+        convert_forward(model.vision_model, vpm_module.Attention, vision_attention_forward)
 
     return model

python/llm/src/ipex_llm/transformers/models/baichuan_m1.py

@@ -0,0 +1,240 @@
+#
+# Copyright 2016 The BigDL Authors.
+#
+# 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
+#
+#     http://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.
+
+# This file is adapted from
+# https://huggingface.co/baichuan-inc/Baichuan-M1-14B-Instruct/blob/main/modeling_baichuan.py
+
+
+import math
+import torch
+import torch.nn.functional as F
+
+from typing import Optional, Tuple, Union
+from transformers.cache_utils import Cache
+from transformers.modeling_outputs import BaseModelOutputWithPast
+from ipex_llm.utils.common import invalidInputError
+from ipex_llm.transformers.models.utils import should_use_fuse_rope, repeat_kv
+from ipex_llm.transformers.models.common import attention_softmax
+from ipex_llm.transformers.models.common import scaled_dot_product_attention
+from ipex_llm.transformers.kv import DynamicNormalCache
+
+
+def pre_register_inv_freq(module: torch.nn.Module):
+    if module.__class__.__name__ == "RotaryEmbedding":
+        inv_freq = module.inv_freq
+        del module.inv_freq
+        module.register_buffer("inv_freq", inv_freq, persistent=False)
+
+
+# copied from Baichuan M1
+def custom_convolution(U, K):
+    """
+    U: Input matrix, shape (bs, seq, h, d)
+    K: Convolution kernel, shape (w, h)
+    Returns: Output matrix V, shape (bs, seq, h, d)
+    """
+    # h, w = K.shape
+    w = K.size(-1)
+    padding = (w - 1, 0)
+    U_padded = F.pad(U, (0, 0, 0, 0, *padding))  # Shape becomes (bs, seq+w-1, h, d)
+    U_unfolded = U_padded.unfold(1, w, 1)  # Shape becomes (bs, seq+w-1, h, d, w)
+    V_unfolded = U_unfolded * K  # Shape remains (bs, seq, h, d, w)
+    V = V_unfolded.sum(dim=-1)  # Shape becomes (bs, seq, h, d)
+    return V
+
+
+def model_forward(
+    self,
+    input_ids: torch.LongTensor = None,
+    attention_mask: Optional[torch.Tensor] = None,
+    position_ids: Optional[torch.LongTensor] = None,
+    seqlens: Optional[torch.LongTensor] = None,
+    past_key_values: Optional[Cache] = None,
+    inputs_embeds: Optional[torch.FloatTensor] = None,
+    use_cache: Optional[bool] = None,
+    output_attentions: Optional[bool] = None,
+    output_hidden_states: Optional[bool] = None,
+    return_dict: Optional[bool] = None,
+    cache_position: Optional[torch.LongTensor] = None,
+) -> Union[Tuple, BaseModelOutputWithPast]:
+    output_attentions = (
+        output_attentions if output_attentions is not None
+        else self.config.output_attentions
+    )
+    output_hidden_states = (
+        output_hidden_states if output_hidden_states is not None
+        else self.config.output_hidden_states
+    )
+    return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+    invalidInputError((input_ids is None) ^ (inputs_embeds is None),
+                      "You cannot specify both input_ids and inputs_embeds at the same time, "
+                      "and must specify either one")
+
+    if inputs_embeds is None:
+        inputs_embeds = self.embed_tokens(input_ids)
+
+    use_cache = use_cache if use_cache is not None else self.config.use_cache
+    use_cache = True if inputs_embeds.device.type == "xpu" else use_cache
+
+    # IPEX-LLM changes start: remove batch multi-pack and use ipex-llm's kv cache
+    # kept for BC (non `Cache` `past_key_values` inputs)
+    if use_cache and not isinstance(past_key_values, DynamicNormalCache):
+        past_key_values = DynamicNormalCache.from_legacy_cache(past_key_values)
+    # IPEX-LLM changes end
+
+    if cache_position is None:
+        past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
+        cache_position = torch.arange(
+            past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1],
+            device=inputs_embeds.device
+        )
+    if position_ids is None:
+        position_ids = cache_position.unsqueeze(0)
+
+    causal_mask = self._update_causal_mask(
+        attention_mask, inputs_embeds, cache_position, past_key_values, output_attentions
+    )
+
+    hidden_states = inputs_embeds
+
+    # create position embeddings to be shared across the decoder layers
+    # position_embeddings = self.rotary_emb(hidden_states, position_ids)
+    position_embeddings = None
+
+    # decoder layers
+    all_hidden_states = () if output_hidden_states else None
+    all_self_attns = () if output_attentions else None
+    next_decoder_cache = None
+
+    for decoder_layer in self.layers:
+        if output_hidden_states:
+            all_hidden_states += (hidden_states,)
+
+        layer_outputs = decoder_layer(
+            hidden_states,
+            attention_mask=causal_mask,
+            position_ids=position_ids,
+            seqlens=None,
+            past_key_value=past_key_values,
+            output_attentions=output_attentions,
+            use_cache=use_cache,
+            cache_position=cache_position,
+            position_embeddings=position_embeddings,
+        )
+
+        hidden_states = layer_outputs[0]
+        if use_cache:
+            next_decoder_cache = layer_outputs[2 if output_attentions else 1]
+
+        if output_attentions:
+            all_self_attns += (layer_outputs[1],)
+
+    hidden_states = self.norm(hidden_states)
+
+    # add hidden states from the last decoder layer
+    if output_hidden_states:
+        all_hidden_states += (hidden_states,)
+
+    next_cache = next_decoder_cache if use_cache else None
+    if not return_dict:
+        return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns]
+                     if v is not None)
+    return BaseModelOutputWithPast(
+        last_hidden_state=hidden_states,
+        past_key_values=next_cache,
+        hidden_states=all_hidden_states,
+        attentions=all_self_attns,
+    )
+
+
+def eager_attention_forward(
+    self,
+    hidden_states: torch.Tensor,
+    attention_mask: Optional[torch.Tensor] = None,
+    position_ids: Optional[torch.LongTensor] = None,
+    seqlens: Optional[torch.LongTensor] = None,
+    past_key_value: Optional[Cache] = None,
+    output_attentions: bool = False,
+    use_cache: bool = False,
+    cache_position: Optional[torch.LongTensor] = None,
+    position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]]=None,
+):
+    invalidInputError(seqlens is None, "`seq_lens` must be None")
+
+    bsz, q_len, _ = hidden_states.size()
+    qkv = self.W_pack(hidden_states)
+    qkv = qkv.view(bsz, q_len, self.num_heads + 2 * self.num_key_value_heads, self.head_dim)
+    query_states, key_states, value_states = qkv.split([self.num_heads,
+                                                        self.num_key_value_heads,
+                                                        self.num_key_value_heads], dim=2)
+    # q, k, v: [bsz, seq_len, num_heads, head_dim]
+
+    if past_key_value is None or past_key_value.get_seq_length(self.layer_idx) == 0:    # prefill
+        self.last_k = key_states[:, -1:]
+        self.last_v = value_states[:, -1:]
+
+        key_states = custom_convolution(key_states, self.conv_k)
+        value_states = custom_convolution(value_states, self.conv_v)
+    else:
+        new_key_states = (self.conv_k[0, 0, :, 0, :1] * self.last_k +
+                          self.conv_k[0, 0, :, 0, 1:] * key_states)
+        self.last_k = key_states
+        key_states = new_key_states
+
+        new_value_states = (self.conv_v[0, 0, :, 0, : 1] * self.last_v +
+                            self.conv_v[0, 0, :, 0, 1:] * value_states)
+        self.last_v = value_states
+        value_states = new_value_states
+
+    query_states = query_states.transpose(1, 2)
+    key_states = key_states.transpose(1, 2)
+    value_states = value_states.transpose(1, 2)
+    # q, k, v: [bsz, num_heads, seq_len, head_dim]
+
+    invalidInputError(should_use_fuse_rope(hidden_states, position_ids, self.training),
+                      "fuse rope must be used")
+    import xe_addons
+    xe_addons.rotary_half_inplaced(self.rotary_emb.inv_freq, position_ids,
+                                   query_states, key_states)
+
+    # ignore sliding window
+    key_states, value_states = past_key_value.update(key_states, value_states,
+                                                     self.layer_idx, None)
+    if self.head_dim <= 128:
+        attn_weights = None
+        attn_output = scaled_dot_product_attention(
+            query_states, key_states, value_states,
+            attention_mask, q_len == key_states.size(2)
+        )
+    else:
+        n_rep = self.num_heads // self.num_key_value_heads
+        key_states = repeat_kv(key_states, n_rep)
+        value_states = repeat_kv(value_states, n_rep)
+        attn_weights = torch.matmul(query_states,
+                                    key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
+        if attention_mask is not None:
+            attn_weights = attn_weights + attention_mask
+        attn_weights = attention_softmax(attn_weights)
+        attn_output = torch.matmul(attn_weights, value_states)
+
+    attn_output = attn_output.transpose(1, 2).contiguous()
+    attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
+
+    attn_output = self.o_proj(attn_output)
+
+    if not output_attentions:
+        attn_weights = None
+    return attn_output, attn_weights, past_key_value