Add supports_gradient_checkpointing

.cache/huggingface/.gitignore

@@ -0,0 +1 @@
+*

.cache/huggingface/download/.gitattributes.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+52373fe24473b1aa44333d318f578ae6bf04b49b
+1734456791.0596795

.cache/huggingface/download/README.md.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+50c19d24e67d8074a25c9eea234d1cf6438c740b
+1734456791.1753678

.cache/huggingface/download/added_tokens.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+dd972e4080e791eab591742c1168ee7fd6279146
+1734456791.1994705

.cache/huggingface/download/config.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+855bd81ff3939cf16c978e81a32b6e56c672e04c
+1734456791.1441224

.cache/huggingface/download/configuration_intern_vit.py.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+7e630c456eb9cf350e55bf850c3ff72f445a7e17
+1734456791.089145

.cache/huggingface/download/configuration_internvl_chat.py.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+799209432caf749e77de4c889ec03fe6a32fcdf9
+1734456791.1727533

.cache/huggingface/download/conversation.py.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+5a771766f21ce3aeeb99b286fb8d188b0038a547
+1734456791.044592

.cache/huggingface/download/generation_config.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+ad56757950bfa85900de6f41f11ef4cca19c6c93
+1734456791.1714978

.cache/huggingface/download/inputs_stats.pth.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+46db3d6cad81aacccef721ea2b5ab612e8db4fc985dad930e7264cd8ff21f3c2
+1734456794.4679294

.cache/huggingface/download/merges.txt.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+31349551d90c7606f325fe0f11bbb8bd5fa0d7c7
+1734456794.464309

.cache/huggingface/download/modeling_intern_vit.py.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+1c5c043a4b860720b3b6e55107e8e6ecf0c573de
+1734456791.741343

.cache/huggingface/download/modeling_internvl_chat.py.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+41f48cd5cb907e025725fc42ac819cf3f03c01b5
+1734456791.6780381

.cache/huggingface/download/outputs_stats.pth.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+0a1e5ab3e5bc804a219212405819a71d42e5f5d74459e16fa6d566f8a85a13cc
+1734456800.2738512

.cache/huggingface/download/preprocessor_config.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+dfd7e50d9d4e67cd679b16b337b419a0c6cfa849
+1734456791.7072408

.cache/huggingface/download/pytorch_model-00001-of-00002.bin.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+483e318a80579169bef059579ff3dc6537a00d2687273d2e2ea930b9b256ab1d
+1734465176.732157

.cache/huggingface/download/pytorch_model-00002-of-00002.bin.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+67802cbbb66929b00889f9f45e996cfbbf6c2a54e710d3b2edd248f3433ddb83
+1734497137.5813448

.cache/huggingface/download/pytorch_model.bin.index.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+a541a2170f4d2ec99333725e52addf2920a53f9e
+1734456793.2043414

.cache/huggingface/download/special_tokens_map.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+74ea6ef5bd52e11d45791fbab7cca3dc984942fa
+1734456793.195179

.cache/huggingface/download/tokenizer.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+34a2790c1c37a3f4774fef44480b2b50e3c0f40f2122d26e057f249460b8735d
+1734456811.4525309

.cache/huggingface/download/tokenizer_config.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+a8e0f5b3e9c57572779dbead982ccd39aa8e8960
+1734456794.3733914

.cache/huggingface/download/vocab.json.metadata

@@ -0,0 +1,3 @@
+6d7aff90716e34e40b62b1d20d52134a2a390211
+4783fe10ac3adce15ac8f358ef5462739852c569
+1734456796.272894

configuration_internvl_chat.py

@@ -63,6 +63,8 @@ class InternVLChatConfig(PretrainedConfig):
         self.ps_version = ps_version  # pixel shuffle version
         self.min_dynamic_patch = min_dynamic_patch
         self.max_dynamic_patch = max_dynamic_patch
+        # By default, we use tie_word_embeddings=False for models of all sizes.
+        self.tie_word_embeddings = self.llm_config.tie_word_embeddings
 
         logger.info(f'vision_select_layer: {self.select_layer}')
         logger.info(f'ps_version: {self.ps_version}')

modeling_intern_vit.py

@@ -364,6 +364,7 @@ class InternVisionEncoder(nn.Module):
 class InternVisionModel(PreTrainedModel):
     main_input_name = 'pixel_values'
     _supports_flash_attn_2 = True
+    supports_gradient_checkpointing = True
     config_class = InternVisionConfig
     _no_split_modules = ['InternVisionEncoderLayer']
 

modeling_internvl_chat.py

@@ -37,6 +37,7 @@ class InternVLChatModel(PreTrainedModel):
     main_input_name = 'pixel_values'
     base_model_prefix = 'language_model'
     _supports_flash_attn_2 = True
+    supports_gradient_checkpointing = True
     _no_split_modules = ['InternVisionModel', 'LlamaDecoderLayer', 'Qwen2DecoderLayer']
 
     def __init__(self, config: InternVLChatConfig, vision_model=None, language_model=None, use_flash_attn=True):
@@ -346,3 +347,13 @@ class InternVLChatModel(PreTrainedModel):
         )
 
         return outputs
+
+    @property
+    def lm_head(self):
+        return self.language_model.get_output_embeddings()
+
+    def get_input_embeddings(self):
+        return self.language_model.get_input_embeddings()
+
+    def get_output_embeddings(self):
+        return self.language_model.get_output_embeddings()