MCPcopy Create free account

hub / github.com/EmbodiedGPT/EmbodiedGPT_Pytorch / functions

Functions424 in github.com/EmbodiedGPT/EmbodiedGPT_Pytorch

⨍ Functions 424◇ Types & classes 107
↓ 20 callersMethodfrom_pretrained
(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs)
robohusky/configuration_husky.py:111
↓ 19 callersFunctionregister_conv_template
Register a new conversation template.
robohusky/conversation.py:186
↓ 18 callersMethodappend_message
Append a new message.
robohusky/conversation.py:124
↓ 17 callersMethod__init__
(self, config)
robohusky/model/modeling_husky.py:242
↓ 17 callersMethod__init__
(self, config)
robohusky/model/modeling_husky_embody2.py:305
↓ 17 callersMethod__init__
(self, config)
robohusky/model/modeling_husky_embody2_ori.py:242
↓ 15 callersFunctionget_conv_template
Get a conversation template.
robohusky/conversation.py:193
↓ 14 callersMethodcopy
(self)
robohusky/conversation.py:158
↓ 14 callersMethodtranspose_for_scores
(self, x)
robohusky/model/modeling_husky_embody2.py:738
↓ 10 callersMethodgenerate
Overrides `generate` function to be able to use the model as a conditional generator. Args: pixel_values (`torch.Flo
robohusky/model/modeling_husky.py:1720
↓ 7 callersFunction_freeze_params
(module)
robohusky/train/train.py:455
↓ 7 callersFunction_freeze_params
(module)
robohusky/train/train_uni.py:454
↓ 7 callersMethodtranspose_for_scores
(self, x)
robohusky/model/modeling_husky.py:662
↓ 7 callersMethodtranspose_for_scores
(self, x)
robohusky/model/modeling_husky_embody2_ori.py:666
↓ 6 callersMethodextract_feature
( self, pixel_values: torch.FloatTensor, )
robohusky/model/modeling_husky.py:1551
↓ 6 callersMethodget_input_embeddings
(self)
robohusky/model/modeling_husky.py:1287
↓ 6 callersMethodget_prompt
Get the prompt for generation.
robohusky/conversation.py:46
↓ 5 callersMethodbatch_decode
This method forwards all its arguments to PreTrainedTokenizer's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstrin
robohusky/model/processing_husky.py:158
↓ 5 callersMethodfrom_pretrained
(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs)
robohusky/model/configuration_husky.py:114
↓ 4 callersMethodask
(self, text, conv, modal_type="image")
demo/script.py:334
↓ 4 callersFunctionget_index
(num_frames, num_segments)
robohusky/video_transformers.py:400
↓ 4 callersMethodget_input_embeddings
(self)
robohusky/model/modeling_husky_embody2.py:1429
↓ 4 callersMethodget_input_embeddings
(self)
robohusky/model/modeling_husky_embody2_ori.py:1288
↓ 4 callersFunctionread_image
(image_file)
demo/script.py:186
↓ 3 callersFunction_prepare_decoder_attention_mask
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
robohusky/llama2_flash_attn_monkey_patch.py:105
↓ 3 callersFunction_prepare_decoder_attention_mask
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
robohusky/train/llama_flash_attn_monkey_patch.py:105
↓ 3 callersMethodanswer
(self, conversations, language_model_inputs, modal_type="image")
demo/script.py:367
↓ 3 callersFunctioncompress_module
(module, target_device)
robohusky/compression.py:52
↓ 3 callersFunctionforward
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,
robohusky/llama2_flash_attn_monkey_patch.py:30
↓ 3 callersFunctionforward
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,
robohusky/train/llama_flash_attn_monkey_patch.py:30
↓ 3 callersMethodglip_show
(self,detect_model,image_glip, object_list,view="left")
demo/script.py:427
↓ 3 callersMethodto_dict
Serializes this instance to a Python dictionary. Override the default [`~PretrainedConfig.to_dict`]. Returns: `Dict[str,
robohusky/configuration_husky.py:311
↓ 3 callersMethodto_dict
Serializes this instance to a Python dictionary. Override the default [`~PretrainedConfig.to_dict`]. Returns: `Dict[
robohusky/model/configuration_husky_ori.py:312
↓ 3 callersMethodto_dict
Serializes this instance to a Python dictionary. Override the default [`~PretrainedConfig.to_dict`]. Returns: `Dict[
robohusky/model/configuration_husky.py:316
↓ 2 callersMethod__init__
(self, vision_config=None, qformer_config=None, text_config=None, num_query_tokens=32, **kwargs)
robohusky/configuration_husky.py:259
↓ 2 callersMethod__init__
(self, vision_config=None, qformer_config=None, text_config=None, num_query_tokens=32, **kwargs)
robohusky/model/configuration_husky_ori.py:260
↓ 2 callersMethod__init__
(self, vision_config=None, qformer_config=None, text_config=None, num_query_tokens=32, **kwargs)
robohusky/model/configuration_husky.py:264
↓ 2 callersFunctionbuild_transform
(input_size)
robohusky/base_dataset.py:41
↓ 2 callersFunctioncompress
Simulate group-wise quantization.
robohusky/compression.py:144
↓ 2 callersMethoddecode
This method forwards all its arguments to PreTrainedTokenizer's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of t
robohusky/model/processing_husky.py:166
↓ 2 callersMethodfill_fix_offset
(more_fix_crop, image_w, image_h, crop_w, crop_h)
robohusky/video_transformers.py:269
↓ 2 callersMethodget_image_embedding
(self, image_file)
demo/script.py:350
↓ 2 callersFunctioninit_dist
(launcher, backend='nccl', **kwargs)
robohusky/dist_utils.py:30
↓ 2 callersFunctionload_image
(image_file, input_size=224)
demo/test.py:126
↓ 2 callersFunctionload_image
(image_file, input_size=224)
demo/inference.py:126
↓ 2 callersFunctionload_video
(video_path, num_segments=8)
demo/test.py:145
↓ 2 callersFunctionload_video
(video_path, num_segments=8)
demo/inference.py:145
↓ 2 callersFunctionreplace_llama_attn_with_flash_attn
()
robohusky/train/llama_flash_attn_monkey_patch.py:128
↓ 2 callersFunctionreplace_llama_rmsnorm_with_fused_rmsnorm
()
robohusky/train/llama_rmsnorm_monkey_patch.py:3
↓ 2 callersFunctionshow_predictions
(scores, boxes, classes)
demo/script.py:86
↓ 2 callersMethodwrite
(self, buf)
robohusky/utils.py:86
↓ 1 callersMethod__init__
(self, dataset, processor, image_path="", input_size=224)
robohusky/base_dataset.py:143
↓ 1 callersMethod__init__
( self, dataset, processor, image_path="", input_s
robohusky/base_dataset_uni.py:241
↓ 1 callersMethod__init__
( self, model_path, device, num_gpus=1, load_
demo/test.py:244
↓ 1 callersMethod__init__
( self, model_path, device, num_gpus=1, load_8bit=
demo/script.py:248
↓ 1 callersMethod__init__
( self, model_path, device, num_gpus=1, load_
demo/inference.py:244
↓ 1 callersFunction_find_free_port
()
robohusky/dist_utils.py:12
↓ 1 callersFunction_init_dist_mpi
(backend, **kwargs)
robohusky/dist_utils.py:51
↓ 1 callersFunction_init_dist_pytorch
(backend, **kwargs)
robohusky/dist_utils.py:43
↓ 1 callersFunction_init_dist_slurm
Initialize slurm distributed training environment. If argument ``port`` is not specified, then the master port will be system environment
robohusky/dist_utils.py:64
↓ 1 callersMethod_init_memcached
(self)
robohusky/base_dataset.py:197
↓ 1 callersFunction_is_free_port
(port)
robohusky/dist_utils.py:23
↓ 1 callersMethod_preprocess_accelerate
r""" Some pre-processing hacks to make the model `accelerate` compatible. Check https://github.com/huggingface/transformers/pull/217
robohusky/model/modeling_husky.py:1592
↓ 1 callersMethod_preprocess_accelerate
r""" Some pre-processing hacks to make the model `accelerate` compatible. Check https://github.com/huggingface/transformers/pull/217
robohusky/model/modeling_husky_embody2.py:1735
↓ 1 callersMethod_preprocess_accelerate
r""" Some pre-processing hacks to make the model `accelerate` compatible. Check https://github.com/huggingface/transformers/pull/21707
robohusky/model/modeling_husky_embody2_ori.py:1594
↓ 1 callersMethod_sample_crop_size
(self, im_size)
robohusky/video_transformers.py:234
↓ 1 callersMethod_sample_fix_offset
(self, image_w, image_h, crop_w, crop_h)
robohusky/video_transformers.py:263
↓ 1 callersMethodanswer
(self, conversations, language_model_inputs, modal_type="image")
demo/test.py:320
↓ 1 callersMethodanswer
(self, conversations, language_model_inputs, modal_type="image")
demo/inference.py:319
↓ 1 callersFunctionapply_compressed_weight
(module, compressed_state_dict, target_device, prefix="")
robohusky/compression.py:81
↓ 1 callersFunctionapply_rotary_pos_emb
(q, k, cos_sin, position_ids)
robohusky/llama2_flash_attn_monkey_patch.py:17
↓ 1 callersFunctionapply_rotary_pos_emb
(q, k, cos_sin, position_ids)
robohusky/train/llama_flash_attn_monkey_patch.py:17
↓ 1 callersMethodask
(self, text, conv, modal_type="image")
demo/test.py:287
↓ 1 callersMethodask
(self, text, conv, modal_type="image")
demo/inference.py:286
↓ 1 callersMethodask_question
(self,input_imgs,detect_model)
demo/script.py:444
↓ 1 callersFunctionbuild_transform
(input_size, norm_type="openai", media_type="image")
robohusky/base_dataset_uni.py:86
↓ 1 callersFunctionconvert_fp16
(in_checkpoint, out_checkpoint)
robohusky/convert_reward_fp16.py:11
↓ 1 callersFunctionconvert_fp16
(in_checkpoint, out_checkpoint)
robohusky/convert_fp16.py:11
↓ 1 callersFunctionconvert_fp16
(in_checkpoint, out_checkpoint)
robohusky/convert_husky_fp16.py:12
↓ 1 callersFunctiondecompress
Simulate group-wise dequantization.
robohusky/compression.py:197
↓ 1 callersFunctionformat_inputs
(sources)
robohusky/base_dataset.py:50
↓ 1 callersFunctionformat_inputs
(sources, conv_tempt="husky", num_query_tokens=256)
robohusky/base_dataset_uni.py:138
↓ 1 callersMethodforward
(self, input: Tensor)
robohusky/compression.py:45
↓ 1 callersMethodforward
(self, hidden_states: torch.Tensor)
robohusky/model/modeling_husky.py:249
↓ 1 callersFunctionget_compressed_list
(module, prefix="")
robohusky/compression.py:65
↓ 1 callersMethodget_decoder
(self)
robohusky/model/modeling_husky.py:1302
↓ 1 callersMethodget_decoder
(self)
robohusky/model/modeling_husky_embody2.py:1444
↓ 1 callersMethodget_decoder
(self)
robohusky/model/modeling_husky_embody2_ori.py:1303
↓ 1 callersMethodget_encoder
(self)
robohusky/model/modeling_husky.py:1299
↓ 1 callersMethodget_encoder
(self)
robohusky/model/modeling_husky_embody2.py:1441
↓ 1 callersMethodget_encoder
(self)
robohusky/model/modeling_husky_embody2_ori.py:1300
↓ 1 callersMethodget_extended_attention_mask
Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_ma
robohusky/model/modeling_husky.py:1073
↓ 1 callersMethodget_extended_attention_mask
Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_ma
robohusky/model/modeling_husky_embody2.py:1218
↓ 1 callersMethodget_extended_attention_mask
Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_mask (
robohusky/model/modeling_husky_embody2_ori.py:1077
↓ 1 callersFunctionget_gpu_memory
(max_gpus=None)
demo/test.py:54
↓ 1 callersFunctionget_gpu_memory
(max_gpus=None)
demo/script.py:114
↓ 1 callersFunctionget_gpu_memory
(max_gpus=None)
demo/inference.py:54
↓ 1 callersMethodget_image_embedding
(self, image_file)
demo/test.py:303
↓ 1 callersMethodget_image_embedding
(self, image_file)
demo/inference.py:302
↓ 1 callersFunctionget_next
()
robohusky/utils.py:211
next →1–100 of 424, ranked by callers