Add RMSNorm unit test (#10190)

.github/workflows/llm_unit_tests.yml

@@ -245,7 +245,6 @@ jobs:
           echo "FALCON_7B_ORIGIN_PATH=${ORIGIN_DIR}/falcon-7b-instruct-with-patch" >> "$GITHUB_ENV"
           echo "MPT_7B_ORIGIN_PATH=${ORIGIN_DIR}/mpt-7b-chat" >> "$GITHUB_ENV"
           echo "WHISPER_TINY_ORIGIN_PATH=${ORIGIN_DIR}/whisper-tiny" >> "$GITHUB_ENV"
-
           echo "MISTRAL_7B_INSTRUCT_V0_1_ORIGIN_PATH=${ORIGIN_DIR}/Mistral-7B-Instruct-v0.1" >> "$GITHUB_ENV"
           echo "BAICHUAN2_7B_ORIGIN_PATH=${ORIGIN_DIR}/Baichuan2-7B-Chat" >> "$GITHUB_ENV"
           echo "QWEN_7B_ORIGIN_PATH=${ORIGIN_DIR}/Qwen-7B-Chat" >> "$GITHUB_ENV"

python/llm/test/inference_gpu/test_transformers_api.py

@@ -34,6 +34,9 @@ print(f'Running on {device}')
     (AutoModel, AutoTokenizer, os.environ.get('CHATGLM2_6B_ORIGIN_PATH')),
     (AutoModelForCausalLM, AutoTokenizer, os.environ.get('FALCON_7B_ORIGIN_PATH')),
     (AutoModelForCausalLM, AutoTokenizer, os.environ.get('MPT_7B_ORIGIN_PATH')),
+    # (AutoModelForCausalLM, AutoTokenizer, os.environ.get('MISTRAL_7B_INSTRUCT_V0_1_ORIGIN_PATH')),
+    # (AutoModelForCausalLM, AutoTokenizer, os.environ.get('BAICHUAN2_7B_ORIGIN_PATH')),
+    # (AutoModelForCausalLM, AutoTokenizer, os.environ.get('QWEN_7B_ORIGIN_PATH')),
     ])
 def test_completion(Model, Tokenizer, model_path, prompt, answer):
     with torch.inference_mode():

python/llm/test/inference_gpu/test_transformers_api_RMSNorm.py

@@ -0,0 +1,154 @@
+
+#
+# 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.
+#
+
+import os
+import gc
+import pytest
+
+import torch
+from bigdl.llm.transformers import AutoModelForCausalLM, AutoModel
+from transformers import LlamaTokenizer, AutoTokenizer
+
+device = os.environ['DEVICE']
+print(f'Running on {device}')
+
+PROMPT = "Once upon a time, there existed a little girl who liked to have adventures. She wanted to go to places and meet new people, and have fun"
+TEST_MODEL_LIST = [
+    ("Llama2-7B", AutoModelForCausalLM, LlamaTokenizer, os.environ.get('LLAMA2_7B_ORIGIN_PATH')),
+    ("ChatGLM2-6B", AutoModel, AutoTokenizer, os.environ.get('CHATGLM2_6B_ORIGIN_PATH')),
+    ("Mistral-7B-Instruct-v0.1", AutoModelForCausalLM, AutoTokenizer, os.environ.get('MISTRAL_7B_INSTRUCT_V0_1_ORIGIN_PATH')),
+    ("Baichuan2-7B-Chat", AutoModelForCausalLM, AutoTokenizer, os.environ.get('BAICHUAN2_7B_ORIGIN_PATH')),
+    ("Qwen-7B-Chat", AutoModelForCausalLM, AutoTokenizer, os.environ.get('QWEN_7B_ORIGIN_PATH')),
+]
+
+class Test_Optimize_Gpu_Model:
+    def setup_method(self):
+        self.layer_outputs = []
+        self.pre_layer_outputs = []
+
+    def run_optimize_gpu_model(self, Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound):
+        with torch.inference_mode():
+            def pre_forward_hook(module, input, output, layer_name):
+                self.pre_layer_outputs.append(output)
+
+            def forward_hook(module, input, output, layer_name):
+                self.layer_outputs.append(output)
+
+
+            tokenizer = Tokenizer.from_pretrained(model_path, trust_remote_code=True)
+            input_ids = tokenizer.encode(PROMPT, return_tensors="pt").to(device)
+
+            model = Model.from_pretrained(model_path,
+                                        load_in_4bit=True,
+                                        optimize_model=False,
+                                        trust_remote_code=True)
+
+            model = model.to(device)
+
+            for layer_name, layer_module in model.named_modules():
+                if layer_name == layer_before_RMSNorm:
+                    layer_module.register_forward_hook(
+                        lambda module, input, output, layer_name=layer_name: pre_forward_hook(module, input,
+                                                                                            output, layer_name))
+                if layer_name == RMSNorm_layer:
+                    layer_module.register_forward_hook(
+                        lambda module, input, output, layer_name=layer_name: forward_hook(module, input,
+                                                                                        output, layer_name))
+            logits_base_model = (model(input_ids)).logits
+            # the list `layer_output` has only one element.
+            layer_tensor = self.layer_outputs.pop()
+            model.to('cpu')
+            opt_model = Model.from_pretrained(model_path,
+                                            load_in_4bit=True,
+                                            optimize_model=True,
+                                            trust_remote_code=True)
+            opt_model = opt_model.to(device)
+
+
+            def replace_forward_hook(module, input, output, layer_name):
+                output = self.pre_layer_outputs[0]
+                return output
+
+            for layer_name, layer_module in opt_model.named_modules():
+                if layer_name == layer_before_RMSNorm:
+                    layer_module.register_forward_hook(
+                        lambda module, input, output, layer_name=layer_name: replace_forward_hook(module, input,
+                                                                                                output, layer_name))
+                if layer_name == RMSNorm_layer:
+                    layer_module.register_forward_hook(
+                        lambda module, input, output, layer_name=layer_name: forward_hook(module, input,
+                                                                                        output, layer_name))
+            logits_optimized_model = (opt_model(input_ids)).logits
+            # the list `layer_output` has only one element.
+            opt_layer_tensor = self.layer_outputs[0]
+            opt_model.to('cpu')
+
+            RMSNorm_output_diff = []
+            for i, (t1, t2) in enumerate(zip(layer_tensor, opt_layer_tensor)):
+                if t1 is not None and t2 is not None:
+                    if isinstance(t1, torch.Tensor) and isinstance(t2, torch.Tensor):
+                        RMSNorm_output_diff.append(t1 - t2)
+            max_diff_tensor = [torch.max(item).item() for item in RMSNorm_output_diff]
+            print(max_diff_tensor)
+            torch.xpu.empty_cache()
+            del model
+            del opt_model
+            gc.collect()
+            assert all(max_diff <= lower_bound for max_diff in max_diff_tensor)
+    
+    @pytest.mark.parametrize('Name, Model, Tokenizer, model_path',TEST_MODEL_LIST)
+    def test_dynamic_functions(self, Name, Model, Tokenizer, model_path):
+        if Name == "Llama2-7B":
+            self.Llama2_7B_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "ChatGLM2-6B":
+            self.Chatglm2_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Mistral-7B-Instruct-v0.1":
+            self.Mistral_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Baichuan2-7B-Chat":
+            self.Baichuan_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Qwen-7B-Chat":
+            self.Qwen_gpu_model(Name, Model, Tokenizer, model_path)
+
+    def Llama2_7B_gpu_model(self, Name, Model, Tokenizer, model_path):
+        layer_before_RMSNorm = "model.layers.30"
+        RMSNorm_layer = "model.layers.31.input_layernorm"
+        lower_bound = 1e-6
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound)
+    
+    def Chatglm2_gpu_model(self, Name, Model, Tokenizer, model_path):
+        layer_before_RMSNorm = "transformer.encoder.layers.26"
+        RMSNorm_layer = "transformer.encoder.layers.27.input_layernorm"
+        lower_bound = 2e-6
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound)
+
+    def Mistral_gpu_model(self, Name, Model, Tokenizer, model_path):
+        layer_before_RMSNorm = "model.layers.30"
+        RMSNorm_layer = "model.layers.31.input_layernorm"
+        lower_bound = 6e-6
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound)
+    
+    def Baichuan_gpu_model(self, Name, Model, Tokenizer, model_path):
+        layer_before_RMSNorm = "model.layers.30"
+        RMSNorm_layer = "model.layers.31.input_layernorm"
+        lower_bound = 5e-7
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound)
+
+    def Qwen_gpu_model(self, Name, Model, Tokenizer, model_path):
+        layer_before_RMSNorm = "transformer.h.30"
+        RMSNorm_layer = "transformer.h.31.ln_1"
+        lower_bound = 2e-6
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, RMSNorm_layer, layer_before_RMSNorm, lower_bound)

python/llm/test/inference_gpu/test_transformers_api_attention.py

@@ -16,6 +16,7 @@
 #
 
 import os
+import gc
 import pytest
 
 import torch
@@ -31,6 +32,9 @@ TEST_MODEL_LIST = [
     ("Llama2-7B", AutoModelForCausalLM, LlamaTokenizer, os.environ.get('LLAMA2_7B_ORIGIN_PATH')),
     ("Falcon-7B", AutoModelForCausalLM, AutoTokenizer, os.environ.get('FALCON_7B_ORIGIN_PATH')),
     ("ChatGLM2-6B", AutoModel, AutoTokenizer, os.environ.get('CHATGLM2_6B_ORIGIN_PATH')),
+    ("Mistral-7B-Instruct-v0.1", AutoModelForCausalLM, AutoTokenizer, os.environ.get('MISTRAL_7B_INSTRUCT_V0_1_ORIGIN_PATH')),
+    ("Baichuan2-7B-Chat", AutoModelForCausalLM, AutoTokenizer, os.environ.get('BAICHUAN2_7B_ORIGIN_PATH')),
+    ("Qwen-7B-Chat", AutoModelForCausalLM, AutoTokenizer, os.environ.get('QWEN_7B_ORIGIN_PATH')),
 ]
 
 class Test_Optimize_Gpu_Model:
@@ -113,6 +117,10 @@ class Test_Optimize_Gpu_Model:
 
             max_diff_tensor = [torch.max(item).item() for item in attn_output_diff]
             print(max_diff_tensor)
+            torch.xpu.empty_cache()
+            del model
+            del opt_model
+            gc.collect()
             
             assert all(max_diff <= lower_bound for max_diff in max_diff_tensor)
     
@@ -126,6 +134,12 @@ class Test_Optimize_Gpu_Model:
             self.Falcon_7B_gpu_model(Name, Model, Tokenizer, model_path)
         elif Name == "ChatGLM2-6B":
             self.Chatglm2_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Mistral-7B-Instruct-v0.1":
+            self.Mistral_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Baichuan2-7B-Chat":
+            self.Baichuan_gpu_model(Name, Model, Tokenizer, model_path)
+        elif Name == "Qwen-7B-Chat":
+            self.Qwen_gpu_model(Name, Model, Tokenizer, model_path)
 
     
     def MPT_7B_gpu_model(self, Name, Model, Tokenizer, model_path):
@@ -139,7 +153,7 @@ class Test_Optimize_Gpu_Model:
         # currently only compare the output of the last self-attention layer.
         layer_norm = "model.layers.31.input_layernorm"
         self_attn = "model.layers.31.self_attn"
-        lower_bound = 5e-2
+        lower_bound = 8e-3
         self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, self_attn, layer_norm, lower_bound)
     
     def Falcon_7B_gpu_model(self, Name, Model, Tokenizer, model_path):
@@ -153,5 +167,26 @@ class Test_Optimize_Gpu_Model:
         # currently only need to compare the output of one self-attention layer.
         layer_norm = "transformer.encoder.layers.27.input_layernorm"
         self_attn = "transformer.encoder.layers.27.self_attention"
-        lower_bound = 5e-3
+        lower_bound = 1e-3
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, self_attn, layer_norm, lower_bound)
+
+    def Mistral_gpu_model(self, Name, Model, Tokenizer, model_path):
+        # currently only need to compare the output of one self-attention layer.
+        layer_norm = "model.layers.31.input_layernorm"
+        self_attn = "model.layers.31.self_attn"
+        lower_bound = 9e-3
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, self_attn, layer_norm, lower_bound)
+    
+    def Baichuan_gpu_model(self, Name, Model, Tokenizer, model_path):
+        # currently only need to compare the output of one self-attention layer.
+        layer_norm = "model.layers.31.input_layernorm"
+        self_attn = "model.layers.31.self_attn"
+        lower_bound = 2e-3
+        self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, self_attn, layer_norm, lower_bound)
+
+    def Qwen_gpu_model(self, Name, Model, Tokenizer, model_path):
+        # currently only need to compare the output of one self-attention layer.
+        layer_norm = "transformer.h.31.ln_1"
+        self_attn = "transformer.h.31.attn"
+        lower_bound = 8e-3
         self.run_optimize_gpu_model(Name, Model, Tokenizer, model_path, self_attn, layer_norm, lower_bound)

python/llm/test/inference_gpu/test_transformers_api_final_logits.py

@@ -16,6 +16,7 @@
 
 
 import os
+import gc
 import pytest
 
 import torch
@@ -48,20 +49,23 @@ def test_optimize_model(Name, Model, Tokenizer, model_path):
 
         model.to('cpu')  # deallocate gpu memory
 
-        model = Model.from_pretrained(model_path,
+        opt_model = Model.from_pretrained(model_path,
                                     load_in_4bit=True,
                                     optimize_model=True,
                                     trust_remote_code=True)
-        model = model.to(device)
+        opt_model = opt_model.to(device)
-        logits_optimized_model = (model(input_ids)).logits
+        logits_optimized_model = (opt_model(input_ids)).logits
-        model.to('cpu')
+        opt_model.to('cpu')
 
         tol = 1e-03
         num_false = torch.isclose(logits_optimized_model, logits_base_model, rtol=tol, atol=tol)\
             .flatten().tolist().count(False)
         
         percent_false = num_false / logits_optimized_model.numel()
-
+        torch.xpu.empty_cache()
+        del model
+        del opt_model
+        gc.collect()
         assert percent_false < 1e-02
 
 

python/llm/test/inference_gpu/test_transformers_api_mlp.py

@@ -15,6 +15,7 @@
 #
  
 import os
+import gc
 import pytest
  
 import torch
@@ -103,7 +104,10 @@ class Test_Optimize_Gpu_Model:
  
             max_diff_tensor = [torch.max(item).item() for item in MLP_output_diff]
             print(max_diff_tensor)
-           
+            torch.xpu.empty_cache()
+            del model
+            del opt_model
+            gc.collect()
             assert all(max_diff <= lower_bound for max_diff in max_diff_tensor)
    
     @pytest.mark.parametrize('Name, Model, Tokenizer, model_path',TEST_MODEL_LIST)

python/llm/test/run-llm-inference-tests-gpu-434.sh

@@ -18,7 +18,9 @@ start=$(date "+%s")
 # fi
 # export OMP_NUM_THREADS=$THREAD_NUM
 export BIGDL_LLM_XMX_DISABLED=1
+pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_attention.py -v -s -k "Mistral"
 pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_mlp.py -v -s -k "Mistral"
+pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_RMSNorm.py -v -s -k "Mistral"
 unset BIGDL_LLM_XMX_DISABLED
 
 now=$(date "+%s")

python/llm/test/run-llm-inference-tests-gpu.sh

@@ -20,8 +20,9 @@ start=$(date "+%s")
 pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api.py -v -s
 export BIGDL_LLM_XMX_DISABLED=1
 pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_final_logits.py -v -s
-pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_attention.py -v -s
+pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_attention.py -v -s -k "not Mistral"
 pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_mlp.py -v -s -k "not Mistral"
+pytest ${LLM_INFERENCE_TEST_DIR}/test_transformers_api_RMSNorm.py -v -s -k "not Mistral"
 unset BIGDL_LLM_XMX_DISABLED
 
 now=$(date "+%s")