samples for new version

Author: Nghi Bui

codetf/models/causal_lm_models/__init__.py

@@ -81,12 +81,13 @@ def load_model_from_config(model_class, model_config, load_in_8bit=False, load_i
         )
 
     def forward(self, sources, max_length=512):
-        encoding = self.tokenizer(sources, return_tensors='pt').to(self.model.device)
+        encoding = self.tokenizer(sources, return_tensors='pt').to(self.device)
         # input_ids = encoding.input_ids.to(self.device)
         # attention_mask = encoding.attention_mask.to(self.device)
         generated_ids = self.model.generate(**encoding, 
                                             max_length=max_length)
 
+        print(generated_ids)
         predictions = self.tokenizer.batch_decode(generated_ids, truncate_before_pattern=[r"\n\n^#", "^'''", "\n\n\n"])
         return predictions
 

requirements.txt

@@ -1,24 +1,23 @@
 # Automatically generated by https://github.com/damnever/pigar.
 
 accelerate==0.20.3
-datasets==2.12.0
-evaluate==0.4.0
-huggingface-hub==0.14.1
+datasets==2.13.1
+huggingface-hub==0.15.1
 iopath==0.1.10
 nltk==3.8.1
-numpy==1.21.6
+numpy==1.25.0
 omegaconf==2.3.0
-pandas==1.3.5
-peft==0.4.0.dev0
+pandas==2.0.2
+peft==0.3.0
 pyparsing==3.0.7
 PyYAML==6.0
-requests==2.31.0
+requests==2.27.1
 rouge-score==0.1.2
 sacrebleu==2.3.1
-salesforce-codetf==1.0.1.1
-scikit-learn==1.0.2
-torch==1.13.1
-torchvision==0.14.1
+scikit-learn==1.2.2
+torch==2.0.1
+torchvision==0.15.2
 tqdm==4.63.0
 transformers==4.30.2
 tree-sitter==0.20.1
+bitsandbytes==0.39.1

setup.py

@@ -2,25 +2,27 @@
 import platform
 
 install_requires = [
-    "datasets==2.12.0",
-    "evaluate==0.4.0",
-    "huggingface-hub==0.14.1",
-    "iopath==0.1.10",
-    "nltk==3.8.1",
-    "numpy==1.22.0",
-    "omegaconf==2.3.0",
-    "pandas==1.3.5",
-    "pyparsing==3.0.7",
-    "PyYAML==6.0",
-    "requests==2.31.0",
-    "rouge-score==0.1.2",
-    "sacrebleu==2.3.1",
-    "scikit-learn==1.0.2",
-    "torch==1.13.1",
-    "torchvision==0.14.1",
-    "tqdm==4.63.0",
-    "tree-sitter==0.20.1",
-    "bitsandbytes==0.39.0"
+  "accelerate==0.20.3",
+  "datasets==2.13.1",
+  "huggingface-hub==0.15.1",
+  "iopath==0.1.10",
+  "nltk==3.8.1",
+  "numpy==1.25.0",
+  "omegaconf==2.3.0",
+  "pandas==2.0.2",
+  "peft==0.3.0",
+  "pyparsing==3.0.7",
+  "PyYAML==6.0",
+  "requests==2.27.1",
+  "rouge-score==0.1.2",
+  "sacrebleu==2.3.1",
+  "scikit-learn==1.2.2",
+  "torch==2.0.1",
+  "torchvision==0.15.2",
+  "tqdm==4.63.0",
+  "transformers==4.30.2",
+  "tree-sitter==0.20.1",
+  "bitsandbytes==0.39.1"
 ]
 
 DEPENDENCY_LINKS = []

test_inference/test_codegen_nl2code.py

@@ -4,8 +4,8 @@
 from codetf.models import load_model_pipeline
 
 code_generation_model = load_model_pipeline(model_name="causallm", task="pretrained",
-            model_type="codegen-350M-mono", is_eval=True,
+            model_type="codegen-2B-mono", is_eval=True,
             load_in_8bit=True, load_in_4bit=False, weight_sharding=False)
 
-result = code_generation_model.predict(["def print_hello_world():"])
+result = code_generation_model.predict(["# this function prints hello world"])
 print(result)

test_inference/test_codet5plus_nl2code.py

@@ -4,8 +4,8 @@
 from codetf.models import load_model_pipeline
 
 code_generation_model = load_model_pipeline(model_name="codet5", task="pretrained",
-            model_type="plus-770M-python", is_eval=True,
-            load_in_8bit=True, load_in_4bit=False, weight_sharding=False)
+            model_type="plus-2B", is_eval=True,
+            load_in_8bit=False, load_in_4bit=False, weight_sharding=True)
 
-result = code_generation_model.predict(["def print_hello_world():"])
+result = code_generation_model.predict(["def print_hello_world():"], max_length=15)
 print(result)

test_inference/test_starcoder_nl2code.py

@@ -8,7 +8,7 @@
             model_type="starcoder-15.5B", is_eval=True,
             load_in_8bit=True, weight_sharding=False)
 
-prompts = "def print_hello_world():"
+prompts = "# this function prints hello world"
 code_snippets = model.predict([prompts])
 
 print(code_snippets)