gpt 13b ckpt

Author: hehesimida

docs/model_cards/gpt2.md

@@ -321,7 +321,8 @@ GPT2支持文本生成和文本分类两个任务的评测。
                                  --eval_dataset_dir {your_path/wikitext-2.mindrecord} \
                                  --run_mode eval \
                                  --epochs 1
-        # PerplexityMetric: {'PerplexityMetric': {'loss': 3.24, 'PPL': 25.55}
+        # gpt2: PerplexityMetric: {'PerplexityMetric': {'loss': 3.24, 'PPL': 25.55}
+        # gpt2_13b(需替换yaml文件): PerplexityMetric: {'PerplexityMetric': {'loss': 2.35, 'PPL': 10.49}
         ```
 
 - 文本分类：

mindformers/mindformer_book.py

@@ -324,6 +324,8 @@ class MindFormerBook:
                 _PROJECT_PATH, "configs/gpt2/run_gpt2_xl.yaml")),
             ("gpt2_xl_lora", os.path.join(
                 _PROJECT_PATH, "configs/gpt2/run_gpt2_xl_lora.yaml")),
+            ("gpt2_13b", os.path.join(
+                _PROJECT_PATH, "configs/gpt2/run_gpt2_13b.yaml")),
             ("llama_7b", os.path.join(
                 _PROJECT_PATH, "configs/llama/run_llama_7b.yaml")),
             ("llama_13b", os.path.join(
@@ -421,7 +423,8 @@ class MindFormerBook:
             'gpt2',
             'gpt2_lora',
             'gpt2_xl',
-            'gpt2_xl_lora'
+            'gpt2_xl_lora',
+            'gpt2_13b'
         ]),
         ('llama', [
             'llama_7b',
@@ -510,7 +513,8 @@ class MindFormerBook:
             'gpt2',
             'gpt2_lora',
             'gpt2_xl',
-            'gpt2_xl_lora'
+            'gpt2_xl_lora',
+            'gpt2_13b'
         ]),
         ('llama', [
             'llama_7b',
@@ -590,7 +594,8 @@ class MindFormerBook:
             'gpt2',
             'gpt2_lora',
             'gpt2_xl',
-            'gpt2_xl_lora'
+            'gpt2_xl_lora',
+            'gpt2_13b'
         ]),
         ('llama', [
             'llama',
@@ -785,6 +790,10 @@ class MindFormerBook:
          ['https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
           '/XFormer_for_mindspore/gpt2/gpt2_xl_lora.ckpt'
           ]),
+        ('gpt2_13b',
+         ['https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
+          '/XFormer_for_mindspore/gpt2/gpt2_13b.ckpt'
+          ]),
         ('pangualpha_2_6b',
          ['https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
           '/XFormer_for_mindspore/pangualpha/pangualpha_2_6b.ckpt'
@@ -1017,6 +1026,12 @@ class MindFormerBook:
           'https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
           '/XFormer_for_mindspore/gpt2/merges.txt'
           ]),
+        ('gpt2_13b',
+         ['https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
+          '/XFormer_for_mindspore/gpt2/vocab.json',
+          'https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
+          '/XFormer_for_mindspore/gpt2/merges.txt'
+          ]),
         ('pangualpha_2_6b',
          ['https://ascend-repo-modelzoo.obs.cn-east-2.myhuaweicloud.com'
           '/XFormer_for_mindspore/pangualpha/vocab.model'