add ziya prompt template

src/cli_demo.py

@@ -4,14 +4,16 @@
 
 
 import torch
-from utils import ModelArguments, FinetuningArguments, load_pretrained, get_logits_processor
-from transformers import HfArgumentParser
+from utils import (
+    load_pretrained,
+    prepare_infer_args,
+    get_logits_processor
+)
 
 
 def main():
 
-    parser = HfArgumentParser((ModelArguments, FinetuningArguments))
-    model_args, finetuning_args = parser.parse_args_into_dataclasses()
+    model_args, data_args, finetuning_args = prepare_infer_args()
     model_name = "BLOOM" if "bloom" in model_args.model_name_or_path else "LLaMA"
     model, tokenizer = load_pretrained(model_args, finetuning_args)
 
@@ -24,14 +26,26 @@ def main():
 
     model.eval()
 
-    def format_example(query):
+    def format_example_alpaca(query, history):
         prompt = "Below is an instruction that describes a task. "
         prompt += "Write a response that appropriately completes the request.\n"
-        prompt += "Instruction:\nHuman: {}\nAssistant: ".format(query)
+        prompt += "Instruction:\n"
+        for old_query, response in history:
+            prompt += "Human: {}\nAssistant: {}\n".format(old_query, response)
+        prompt += "Human: {}\nAssistant:".format(query)
         return prompt
 
+    def format_example_ziya(query, history):
+        prompt = ""
+        for old_query, response in history:
+            prompt += "<human>: {}\n<bot>: {}\n".format(old_query, response)
+        prompt += "<human>: {}\n<bot>:".format(query)
+        return prompt
+
+    format_example = format_example_alpaca if data_args.prompt_template == "alpaca" else format_example_ziya
+
     def predict(query, history: list):
-        input_ids = tokenizer([format_example(query)], return_tensors="pt")["input_ids"]
+        input_ids = tokenizer([format_example(query, history)], return_tensors="pt")["input_ids"]
         input_ids = input_ids.to(model.device)
         gen_kwargs = {
             "do_sample": True,
@@ -65,6 +79,7 @@ def main():
 
         if query.strip() == "clear":
             history = []
+            print("History has been removed.")
             continue
 
         response, history = predict(query, history)

src/export_model.py

@@ -3,19 +3,15 @@
 # Usage: python export_model.py --checkpoint_dir path_to_checkpoint --output_dir path_to_save_model
 
 
-from transformers import HfArgumentParser, TrainingArguments
-from utils import ModelArguments, FinetuningArguments, load_pretrained
+from utils import load_pretrained, prepare_args
 
 
 def main():
 
-    parser = HfArgumentParser((ModelArguments, TrainingArguments, FinetuningArguments))
-    model_args, training_args, finetuning_args = parser.parse_args_into_dataclasses()
-
+    model_args, _, training_args, finetuning_args = prepare_args(stage="sft")
     model, tokenizer = load_pretrained(model_args, finetuning_args)
     model.save_pretrained(training_args.output_dir, max_shard_size="10GB")
     tokenizer.save_pretrained(training_args.output_dir)
-
     print("model and tokenizer have been saved at:", training_args.output_dir)
 
 

src/utils/__init__.py

@@ -1,6 +1,7 @@
 from .common import (
     load_pretrained,
     prepare_args,
+    prepare_infer_args,
     prepare_data,
     preprocess_data
 )
@@ -13,5 +14,4 @@ from .seq2seq import ComputeMetrics, Seq2SeqPeftTrainer
 from .pairwise import PairwiseDataCollatorWithPadding, PairwisePeftTrainer
 from .ppo import PPOPeftTrainer
 
-from .config import ModelArguments, FinetuningArguments
 from .other import get_logits_processor, plot_loss

src/utils/common.py

@@ -264,6 +264,18 @@ def prepare_args(
     return model_args, data_args, training_args, finetuning_args
 
 
+def prepare_infer_args() -> Tuple[ModelArguments, DataTrainingArguments, FinetuningArguments]:
+
+    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, FinetuningArguments))
+
+    if len(sys.argv) == 2 and sys.argv[1].endswith(".json"): # Provide arguments with a json file.
+        model_args, data_args, finetuning_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
+    else:
+        model_args, data_args, finetuning_args = parser.parse_args_into_dataclasses()
+
+    return model_args, data_args, finetuning_args
+
+
 def prepare_data(
         model_args: ModelArguments,
         data_args: DataTrainingArguments
@@ -347,7 +359,8 @@ def preprocess_data(
     column_names = list(dataset.column_names)
     prefix = data_args.source_prefix if data_args.source_prefix is not None else ""
 
-    def format_example(examples): # support question with a single answer or multiple answers
+    # support question with a single answer or multiple answers
+    def format_example_alpaca(examples):
         for i in range(len(examples["prompt"])):
             if examples["prompt"][i] and examples["response"][i]:
                 query, answer = examples["prompt"][i], examples["response"][i]
@@ -357,12 +370,27 @@ def preprocess_data(
                 prompt += "Write a response that appropriately completes the request.\n"
                 prompt += "Instruction:\n" + prefix
                 if examples["history"][i]:
-                    history = examples["history"][i]
-                    for old_query, response in history:
+                    for old_query, response in examples["history"][i]:
                         prompt += "Human: {}\nAssistant: {}\n".format(old_query, response)
                 prompt += "Human: {}\nAssistant: ".format(query)
                 yield prompt, answer
 
+    def format_example_ziya(examples):
+        for i in range(len(examples["prompt"])):
+            if examples["prompt"][i] and examples["response"][i]:
+                query, answer = examples["prompt"][i], examples["response"][i]
+                if examples["query"][i]:
+                    query += "\n" + examples["query"][i]
+                prompt = ""
+                if examples["history"][i]:
+                    for old_query, response in examples["history"][i]:
+                        prompt += "<human>: {}\n<bot>: {}\n".format(old_query, response)
+                prompt += "<human>: {}\n<bot>:".format(query)
+                prompt = prefix + prompt
+                yield prompt, answer
+
+    format_example = format_example_alpaca if data_args.prompt_template == "alpaca" else format_example_ziya
+
     def preprocess_pretrain_dataset(examples):
         # build grouped texts with format `<s> X1 X2 X3 ...` (without </s>)
         text_ids = tokenizer(examples["prompt"])["input_ids"]

src/utils/config.py

@@ -136,6 +136,10 @@ class DataTrainingArguments:
         default=0,
         metadata={"help": "Proportion of the dataset to include in the development set, should be between 0.0 and 1.0."}
     )
+    prompt_template: Optional[Literal["alpaca", "ziya"]] = field(
+        default="alpaca",
+        metadata={"help": "Which template to use for constructing prompts in training."}
+    )
 
     def __post_init__(self): # support mixing multiple datasets
         dataset_names = [ds.strip() for ds in self.dataset.split(",")]

src/web_demo.py

@@ -7,14 +7,12 @@ import torch
 import mdtex2html
 import gradio as gr
 
-from utils import ModelArguments, FinetuningArguments, load_pretrained, get_logits_processor
-from transformers import HfArgumentParser
+from utils import load_pretrained, prepare_infer_args, get_logits_processor
 from transformers.utils.versions import require_version
 
 
 require_version("gradio==3.27.0", "To fix: pip install gradio==3.27.0") # higher version may cause problems
-parser = HfArgumentParser((ModelArguments, FinetuningArguments))
-model_args, finetuning_args = parser.parse_args_into_dataclasses()
+model_args, data_args, finetuning_args = prepare_infer_args()
 model, tokenizer = load_pretrained(model_args, finetuning_args)
 
 if torch.cuda.device_count() > 1:
@@ -75,17 +73,31 @@ def parse_text(text): # copy from https://github.com/GaiZhenbiao/ChuanhuChatGPT
     return text
 
 
-def format_example(query):
+def format_example_alpaca(query, history):
     prompt = "Below is an instruction that describes a task. "
     prompt += "Write a response that appropriately completes the request.\n"
-    prompt += "Instruction:\nHuman: {}\nAssistant: ".format(query)
+    prompt += "Instruction:\n"
+    for old_query, response in history:
+        prompt += "Human: {}\nAssistant: {}\n".format(old_query, response)
+    prompt += "Human: {}\nAssistant:".format(query)
     return prompt
 
 
+def format_example_ziya(query, history):
+    prompt = ""
+    for old_query, response in history:
+        prompt += "<human>: {}\n<bot>: {}\n".format(old_query, response)
+    prompt += "<human>: {}\n<bot>:".format(query)
+    return prompt
+
+
+format_example = format_example_alpaca if data_args.prompt_template == "alpaca" else format_example_ziya
+
+
 def predict(input, chatbot, max_length, top_p, temperature, history):
     chatbot.append((parse_text(input), ""))
 
-    input_ids = tokenizer([format_example(input)], return_tensors="pt")["input_ids"]
+    input_ids = tokenizer([format_example(input, history)], return_tensors="pt")["input_ids"]
     input_ids = input_ids.to(model.device)
     gen_kwargs = {
         "do_sample": True,