From d81b99541309a803f8e5276abded1910010b083e Mon Sep 17 00:00:00 2001
From: coffee <hanluzhi0429@gmail.com>
Date: Wed, 7 May 2025 11:26:23 +0800
Subject: [PATCH] add dolly test code

---
 test/temp_test.py                    | 20 ++++++++++++++++++++
 test/test_dolly.py                   | 19 +++++++++++++++++++
 test/test_huggingface_train.py       |  0
 test/test_huggingface_transformer.py | 14 ++++++++++++++
 4 files changed, 53 insertions(+)
 create mode 100755 test/temp_test.py
 create mode 100755 test/test_dolly.py
 create mode 100755 test/test_huggingface_train.py
 create mode 100755 test/test_huggingface_transformer.py

diff --git a/test/temp_test.py b/test/temp_test.py
new file mode 100755
index 0000000..0d7ba88
--- /dev/null
+++ b/test/temp_test.py
@@ -0,0 +1,20 @@
+# from transformers import PreTrainedModel
+from transformers.modeling_utils import PreTrainedModel
+import torch
+from torch import nn
+from src_llm.configuration_dolly import DollyConfig
+from src_llm.modeling_dolly import DollyModel
+# model = DollyModel(DollyConfig())
+# model.save_pretrained("./my_model")
+# model.config.save_pretrained("./my_model")
+
+config = DollyConfig.from_pretrained("./my_model")
+model = DollyModel.from_pretrained("./my_model", config=config)
+
+
+# input_ids = torch.tensor([[101, 1024, 102]]).long() 
+# outputs = model(input_ids)
+# print(outputs)
+
+model_size = sum(t.numel() for t in model.parameters())
+print(f"Model Size: {model_size/1000**2:.1f}M parameters")
\ No newline at end of file
diff --git a/test/test_dolly.py b/test/test_dolly.py
new file mode 100755
index 0000000..c530265
--- /dev/null
+++ b/test/test_dolly.py
@@ -0,0 +1,19 @@
+# from transformers import PreTrainedModel
+from transformers.modeling_utils import PreTrainedModel
+import torch
+from torch import nn
+from src_llm.configuration_dolly import DollyConfig
+from src_llm.modeling_dolly import DollyModel
+#从源码构建dolly模型
+model = DollyModel(DollyConfig())
+model.save_pretrained("weights_llm")
+model.config.save_pretrained("weights_llm")
+model_size = sum(t.numel() for t in model.parameters())
+print(f"Model Size: {model_size/1000**2:.1f}M parameters")
+
+#从配置文件和权重构建dolly模型
+config = DollyConfig.from_pretrained("./my_model")
+model = DollyModel.from_pretrained("./my_model", config=config)
+input_ids = torch.tensor([[101, 1024, 102]]).long() 
+outputs = model(input_ids)
+print(outputs)
diff --git a/test/test_huggingface_train.py b/test/test_huggingface_train.py
new file mode 100755
index 0000000..e69de29
diff --git a/test/test_huggingface_transformer.py b/test/test_huggingface_transformer.py
new file mode 100755
index 0000000..26cbab1
--- /dev/null
+++ b/test/test_huggingface_transformer.py
@@ -0,0 +1,14 @@
+import transformers
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import LlamaForCausalLM
+from transformers import LlamaConfig
+config = LlamaConfig(
+    vocab_size=32000,  
+    hidden_size=4096,  
+    num_attention_heads=32,  
+    num_hidden_layers=32, 
+    intermediate_size=11008,  
+)
+model = LlamaForCausalLM(config)
+model_size = sum(t.numel() for t in model.parameters())
+print(f"Model Size: {model_size/1000**2:.1f}M parameters")
\ No newline at end of file