train: modify inference and add start end time

results/inference/Baichuan2-7B/Baichuan2_predict_1.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_1

results/inference/Baichuan2-7B/Baichuan2_predict_1_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_1_single

results/inference/Baichuan2-7B/Baichuan2_predict_2.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_2

results/inference/Baichuan2-7B/Baichuan2_predict_2_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_2_single

results/inference/Baichuan2-7B/Baichuan2_predict_3.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_3

results/inference/Baichuan2-7B/Baichuan2_predict_3_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Baichuan2-7B/Baichuan2_predict_3_single

results/inference/ChatGLM2-6B/ChatGLM2_predict_1.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_1

results/inference/ChatGLM2-6B/ChatGLM2_predict_1_single.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_1_single

results/inference/ChatGLM2-6B/ChatGLM2_predict_2.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_2

results/inference/ChatGLM2-6B/ChatGLM2_predict_2_single.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_2_single

results/inference/ChatGLM2-6B/ChatGLM2_predict_3.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_3

results/inference/ChatGLM2-6B/ChatGLM2_predict_3_single.yaml

@@ -1,5 +1,5 @@
 ### model
-model_name_or_path: ZhipuAI/chatglm2-6b
+model_name_or_path: ../../llm/chatglm/data
 
 ### method
 do_predict: true
@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/ChatGLM2-6B/ChatGLM2_predict_3_single

results/inference/Llama2-7B/llama2_predict_1.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_1

results/inference/Llama2-7B/llama2_predict_1_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_1_single

results/inference/Llama2-7B/llama2_predict_2.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_2

results/inference/Llama2-7B/llama2_predict_2_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_2_single

results/inference/Llama2-7B/llama2_predict_3.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_3

results/inference/Llama2-7B/llama2_predict_3_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Llama2-7B/llama2_predict_3_single

results/inference/Qwen-7B/Qwen_predict_1.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_1

results/inference/Qwen-7B/Qwen_predict_1_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_1_single

results/inference/Qwen-7B/Qwen_predict_2.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_2

results/inference/Qwen-7B/Qwen_predict_2_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_2_single

results/inference/Qwen-7B/Qwen_predict_3.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_3

results/inference/Qwen-7B/Qwen_predict_3_single.yaml

@@ -11,6 +11,7 @@ cutoff_len: 1024
 max_samples: 50
 overwrite_cache: true
 preprocessing_num_workers: 16
+include_tokens_per_second: true
 
 ### output
 output_dir: ./results/inference/Qwen-7B/Qwen_predict_3_single

src/llamafactory/train/sft/workflow.py

@@ -15,6 +15,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import time
 from typing import TYPE_CHECKING, List, Optional
 
 from ...data import SFTDataCollatorWith4DAttentionMask, get_dataset
@@ -112,7 +113,9 @@ def run_sft(
 
     # Predict
     if training_args.do_predict:
+        print("predict start time: " + time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()))
         predict_results = trainer.predict(dataset_module["eval_dataset"], metric_key_prefix="predict", **gen_kwargs)
+        print("predict end time: " + time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()))
         if training_args.predict_with_generate:  # predict_loss will be wrong if predict_with_generate is enabled
             predict_results.metrics.pop("predict_loss", None)
         trainer.log_metrics("predict", predict_results.metrics)