Support PP inference for chatglm3 (#11375)

Author: plusbang

python/llm/example/GPU/Pipeline-Parallel-Inference/README.md

@@ -12,6 +12,7 @@ To run this example with IPEX-LLM on Intel GPUs, we have some recommended requir
 - [Qwen/Qwen1.5-7B-Chat](./run_qwen1.5_arc_2_card.sh)
 - [Qwen/Qwen1.5-14B-Chat](./run_qwen1.5_arc_2_card.sh)
 - [Qwen/Qwen1.5-32B-Chat](./run_qwen1.5_arc_2_card.sh)
+- [THUDM/chatglm3-6b](./run_chatglm_arc_2_card.sh)
 - [baichuan-inc/Baichuan2-7B-Chat](./run_baichuan2_arc_2_card.sh)
 - [baichuan-inc/Baichuan2-13B-Chat](./run_baichuan2_arc_2_card.sh)
 - [microsoft/Phi-3-mini-4k-instruct](./run_phi3_arc_2_card.sh)
@@ -71,6 +72,21 @@ bash run_qwen1.5_arc_2_card.sh
 
 </details>
 
+<details>
+  <summary> Show chatglm example </summary>
+
+#### Run chatglm3-6B on two Intel Arc A770
+
+You could specify `--repo-id-or-model-path` in the test script to be the huggingface repo id for chatglm to be downloaded, or the path to the huggingface checkpoint folder. Besides, you could change `NUM_GPUS` to the number of GPUs you have on your machine.
+
+```bash
+bash run_chatglm_arc_2_card.sh
+```
+
+</details>
+
+</details>
+
 <details>
   <summary> Show Baichuan2 example </summary>
 

python/llm/example/GPU/Pipeline-Parallel-Inference/generate.py

@@ -19,7 +19,7 @@
 import time
 import argparse
 
-from ipex_llm.transformers import AutoModelForCausalLM, init_pipeline_parallel
+from ipex_llm.transformers import AutoModel, AutoModelForCausalLM, init_pipeline_parallel
 from transformers import AutoTokenizer
 
 init_pipeline_parallel()
@@ -41,13 +41,21 @@
 
     # Load model in 4 bit,
     # which convert the relevant layers in the model into INT4 format
-    model = AutoModelForCausalLM.from_pretrained(model_path,
-                                                 load_in_4bit=True,
-                                                 optimize_model=True,
-                                                 trust_remote_code=True,
-                                                 use_cache=True,
-                                                 torch_dtype=torch.float16,
-                                                 pipeline_parallel_stages=args.gpu_num)
+    try:
+        model = AutoModelForCausalLM.from_pretrained(model_path,
+                                                     load_in_4bit=True,
+                                                     optimize_model=True,
+                                                     trust_remote_code=True,
+                                                     use_cache=True,
+                                                     torch_dtype=torch.float16,
+                                                     pipeline_parallel_stages=args.gpu_num)
+    except:
+        model = AutoModel.from_pretrained(model_path,
+                                          load_in_4bit=True,
+                                          optimize_model=True,
+                                          trust_remote_code=True,
+                                          use_cache=True,
+                                          pipeline_parallel_stages=args.gpu_num)
 
     # Load tokenizer
     tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)

python/llm/example/GPU/Pipeline-Parallel-Inference/run_chatglm_arc_2_card.sh

@@ -0,0 +1,31 @@
+#
+# 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.
+#
+
+source /opt/intel/oneapi/setvars.sh
+export MASTER_ADDR=127.0.0.1
+export MASTER_PORT=9090
+export FI_PROVIDER=tcp
+export USE_XETLA=OFF
+export OMP_NUM_THREADS=6
+if [[ $KERNEL_VERSION != *"6.5"* ]]; then
+    export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
+fi
+export TORCH_LLM_ALLREDUCE=0
+
+NUM_GPUS=2 # number of used GPU
+# To run chatglm3-6b
+CCL_ZE_IPC_EXCHANGE=sockets torchrun --standalone --nnodes=1 --nproc-per-node $NUM_GPUS \
+    generate.py --repo-id-or-model-path 'THUDM/chatglm3-6b' --gpu-num $NUM_GPUS

python/llm/src/ipex_llm/transformers/models/chatglm2.py

@@ -74,10 +74,12 @@ def chatglm2_model_forward(
     use_cache = use_cache if use_cache is not None else self.config.use_cache
     return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
-    batch_size, seq_length = input_ids.shape
-
     if inputs_embeds is None:
+        batch_size, seq_length = input_ids.shape
         inputs_embeds = self.embedding(input_ids)
+    else:
+        inputs_embeds = inputs_embeds.transpose(0, 1).contiguous()
+        seq_length, batch_size, _ = inputs_embeds.shape
 
     if full_attention_mask is None:
         if (attention_mask is not None and not attention_mask.all()) or (

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

@@ -71,6 +71,19 @@ def forward(self, hidden_states, past_key_value=None, use_cache=False, **kwargs)
         return outputs
 
 
+class Dummy_GLMBlock(nn.Module):
+    def __init__(self, *args):
+        super().__init__()
+        # to avoid AttributeError
+        self.input_layernorm = DummyLayer()
+        self.mlp = Dummy_MLPLayer()
+
+    def forward(
+            self, hidden_states, attention_mask, rotary_pos_emb, kv_cache=None, use_cache=True,
+    ):
+        return hidden_states, kv_cache
+
+
 def init_pipeline_parallel():
     import oneccl_bindings_for_pytorch
     os.environ["MASTER_ADDR"] = os.environ.get("MASTER_ADDR", "127.0.0.1")
@@ -79,28 +92,49 @@ def init_pipeline_parallel():
 
 
 def pipeline_parallel(model, pipeline_parallel_stages):
-    slice_size = (model.config.num_hidden_layers + pipeline_parallel_stages - 1) // \
-        pipeline_parallel_stages
+    global num_layers
+    if hasattr(model.config, 'num_hidden_layers'):
+        num_layers = model.config.num_hidden_layers
+    elif hasattr(model.config, 'num_layers'):
+        # for chatglm3-6b
+        num_layers = model.config.num_layers
+
+    slice_size = (num_layers + pipeline_parallel_stages - 1) // pipeline_parallel_stages
 
     local_rank = dist.get_rank()
 
     global layer_start
     global layer_end
     layer_start = slice_size * local_rank
-    layer_end = layer_start + min(slice_size, model.config.num_hidden_layers - layer_start)
-
-    for i in range(model.config.num_hidden_layers):
-        if i < layer_start or i >= layer_end:
-            model._modules['model'].layers[i] = Dummy_DecoderLayer()
-        else:
-            # align layer_idx and len(past_key_values), otherwise abnormal output
-            model._modules['model'].layers[i].self_attn.layer_idx = i - layer_start
-
-    if local_rank != 0:
-        model._modules['model'].embed_tokens = DummyLayer()
-    if local_rank != pipeline_parallel_stages - 1:
-        model._modules['model'].norm = DummyLayer()
-        model._modules['lm_head'] = DummyLayer()
+    layer_end = layer_start + min(slice_size, num_layers - layer_start)
+
+    if model.config.architectures is not None \
+       and model.config.architectures[0] in ["ChatGLMModel", "ChatGLMForConditionalGeneration"]:
+        # for chatglm3-6b
+        for i in range(num_layers):
+            if i < layer_start or i >= layer_end:
+                model._modules['transformer'].encoder.layers[i] = Dummy_GLMBlock()
+            else:
+                model._modules['transformer'].encoder.layers[i].self_attention.num_layers = \
+                    i - layer_start
+
+        if local_rank != 0:
+            model._modules['transformer'].embedding = DummyLayer()
+        if local_rank != pipeline_parallel_stages - 1:
+            model._modules['transformer'].encoder.final_layernorm = DummyLayer()
+            model._modules['transformer'].output_layer = DummyLayer()
+    else:
+        for i in range(num_layers):
+            if i < layer_start or i >= layer_end:
+                model._modules['model'].layers[i] = Dummy_DecoderLayer()
+            else:
+                model._modules['model'].layers[i].self_attn.layer_idx = i - layer_start
+
+        if local_rank != 0:
+            model._modules['model'].embed_tokens = DummyLayer()
+        if local_rank != pipeline_parallel_stages - 1:
+            model._modules['model'].norm = DummyLayer()
+            model._modules['lm_head'] = DummyLayer()
 
     model.pipeline_parallel_stages = pipeline_parallel_stages
     model = model.to(f'xpu:{local_rank}')
@@ -176,6 +210,7 @@ def pipeline_parallel_generate(self,
 
     global layer_start
     global layer_end
+    global num_layers
 
     self.first_token_time = 0
     self.next_token_time = []