Feature/chatglm (#4240)

* [shardformer] added tests

* [shardformer] vit test finish and support

* [shardformer] chatglm ready

* import chatglm

* [shardformer] add test kit in model zoo for chatglm

* [sharformer] add first version of policy of chatglm

* [shardformer] polish chatglm code

* [shardformer] polish code

* [shardformer] support chatglm without layernorm

* [shardformer] chatglm shard without mlp sharding

* [shardformer] delete some file

* [shardformer] ChatGLM support layernorm sharding

* [shardformer] register without auto policy

* [shardformer] pre-commit check files

* [shardformer] fix chatglm configuration with pre-commit

tests/kit/model_zoo/transformers/__init__.py

@@ -1,6 +1,7 @@
 from .albert import *
 from .bert import *
 from .bloom import *
+from .chatglm import *
 from .gpt import *
 from .llama import *
 from .opt import *

tests/kit/model_zoo/transformers/chatglm.py

@@ -0,0 +1,38 @@
+import torch
+import transformers
+
+from ..registry import ModelAttribute, model_zoo
+from .chatglm2_6b.configuration_chatglm import ChatGLMConfig
+from .chatglm2_6b.modeling_chatglm import ChatGLMModel
+
+# ================================
+# Register single-sentence ChatGLM
+# ================================
+
+
+def data_gen():
+    input_ids = torch.tensor([[5941, 15, 2670, 3543, 632, 2075]], dtype=torch.int64)
+    attention_mask = torch.tensor([[1, 1, 1, 1, 1, 1]])
+    return dict(input_ids=input_ids, attention_mask=attention_mask)
+
+
+# define output transform function
+output_transform_fn = lambda x: x
+
+# define loss function
+loss_fn_for_chatglm_model = lambda x: x.last_hidden_state.mean()
+loss_fn = lambda x: x.loss
+config = ChatGLMConfig(num_layers=1,
+                       padded_vocab_size=65024,
+                       hidden_size=64,
+                       num_attention_heads=8,
+                       rmsnorm=False,
+                       original_rope=True,
+                       use_cache=True)
+
+model_zoo.register(name='transformers_chatglm',
+                   model_fn=lambda: ChatGLMModel(config, empty_init=False),
+                   data_gen_fn=data_gen,
+                   output_transform_fn=output_transform_fn,
+                   loss_fn=loss_fn_for_chatglm_model,
+                   model_attribute=ModelAttribute(has_control_flow=True))

tests/kit/model_zoo/transformers/chatglm2_6b/configuration_chatglm.py

@@ -0,0 +1,58 @@
+from transformers import PretrainedConfig
+
+
+class ChatGLMConfig(PretrainedConfig):
+    model_type = "chatglm"
+
+    def __init__(self,
+                 num_layers=28,
+                 padded_vocab_size=65024,
+                 hidden_size=4096,
+                 ffn_hidden_size=13696,
+                 kv_channels=128,
+                 num_attention_heads=32,
+                 seq_length=2048,
+                 hidden_dropout=0.0,
+                 attention_dropout=0.0,
+                 layernorm_epsilon=1e-5,
+                 rmsnorm=True,
+                 apply_residual_connection_post_layernorm=False,
+                 post_layer_norm=True,
+                 add_bias_linear=False,
+                 add_qkv_bias=False,
+                 bias_dropout_fusion=True,
+                 multi_query_attention=False,
+                 multi_query_group_num=1,
+                 apply_query_key_layer_scaling=True,
+                 attention_softmax_in_fp32=True,
+                 fp32_residual_connection=False,
+                 quantization_bit=0,
+                 pre_seq_len=None,
+                 prefix_projection=False,
+                 **kwargs):
+        self.num_layers = num_layers
+        self.vocab_size = padded_vocab_size
+        self.padded_vocab_size = padded_vocab_size
+        self.hidden_size = hidden_size
+        self.ffn_hidden_size = ffn_hidden_size
+        self.kv_channels = kv_channels
+        self.num_attention_heads = num_attention_heads
+        self.seq_length = seq_length
+        self.hidden_dropout = hidden_dropout
+        self.attention_dropout = attention_dropout
+        self.layernorm_epsilon = layernorm_epsilon
+        self.rmsnorm = rmsnorm
+        self.apply_residual_connection_post_layernorm = apply_residual_connection_post_layernorm
+        self.post_layer_norm = post_layer_norm
+        self.add_bias_linear = add_bias_linear
+        self.add_qkv_bias = add_qkv_bias
+        self.bias_dropout_fusion = bias_dropout_fusion
+        self.multi_query_attention = multi_query_attention
+        self.multi_query_group_num = multi_query_group_num
+        self.apply_query_key_layer_scaling = apply_query_key_layer_scaling
+        self.attention_softmax_in_fp32 = attention_softmax_in_fp32
+        self.fp32_residual_connection = fp32_residual_connection
+        self.quantization_bit = quantization_bit
+        self.pre_seq_len = pre_seq_len
+        self.prefix_projection = prefix_projection
+        super().__init__(**kwargs)