MCPcopy Create free account

hub / github.com/OpenBMB/AgentCPM-GUI / functions

Functions568 in github.com/OpenBMB/AgentCPM-GUI

↓ 16 callersMethodencode
(self, image_paths: List[str])
eval/utils/utils_odyssey/visual.py:416
↓ 14 callersMethodgenerate
( self, inputs: Optional[torch.Tensor] = None, generation_config: Optional[Generati
eval/utils/utils_odyssey/modeling_qwen.py:1219
↓ 9 callersMethodconvert_tokens_to_ids
( self, tokens: Union[bytes, str, List[Union[bytes, str]]] )
eval/utils/utils_odyssey/tokenization_qwen.py:206
↓ 8 callersMethod__init__
(self, config)
eval/utils/utils_odyssey/modeling_qwen.py:462
↓ 8 callersFunctioncompact_json_dumps
(obj)
rft/trainer/utils/dataset.py:233
↓ 8 callersFunctionget_dataset_dir
(data_name)
eval/utils/utils.py:89
↓ 7 callersMethodsave
(self, filepath)
eval/utils/utils_odyssey/tokenization_qwen.py:523
↓ 7 callersMethodstart
启动同步管理器
rft/trainer/zmq.py:129
↓ 5 callersMethod_save
(self, output_dir: Optional[str] = None, state_dict=None)
sft/trainer.py:216
↓ 5 callersFunctionis_tap_action
(normalized_start_yx, normalized_end_yx)
eval/utils/action_utils.py:94
↓ 4 callersMethod__init__
( self, grid_size, embed_dim, num_heads, kv_d
eval/utils/utils_odyssey/visual.py:99
↓ 4 callersFunction_tokenize_str
(role, content)
eval/utils/utils_odyssey/qwen_generation_utils.py:136
↓ 4 callersFunctionbuild_transform
(data_args)
sft/finetune.py:147
↓ 4 callersFunctioncalculate_dist_score
(pred_loc: list[list[int,int]], gt_loc: list[int,int], res: tuple[int,int], bbox: list[int])
rft/trainer/utils/gui_eval.py:351
↓ 4 callersMethodchat
( self, tokenizer: PreTrainedTokenizer, query: str, history: Optional[Hist
eval/utils/utils_odyssey/modeling_qwen.py:1096
↓ 4 callersFunctionensure_divide
(length, patch_size)
sft/dataset.py:499
↓ 4 callersFunctionload_and_validate_action
(res:str,)
rft/trainer/utils/gui_eval.py:10
↓ 4 callersMethodlog
(self, logs: dict[str, float], start_time: Optional[float] = None)
rft/trainer/arl.py:783
↓ 4 callersFunctionrank0_print
(*args)
sft/finetune.py:75
↓ 4 callersFunctiontrim_and_pad
(seq, batch_first, padding_value)
sft/dataset.py:105
↓ 3 callersMethod_get_per_token_logps
(self, model, inputs)
rft/trainer/arl.py:689
↓ 3 callersFunction_replace_closed_tag
( input_tokens: List[Any], start_tags: Union[Any, Tuple[Any]], end_tags: Union[Any, Tuple[Any]]
eval/utils/utils_odyssey/tokenization_qwen.py:71
↓ 3 callersMethod_split_heads
(self, tensor, num_heads, attn_head_size)
eval/utils/utils_odyssey/modeling_qwen.py:383
↓ 3 callersFunctionaitw_2_qwen2_5_action
Convert AITW action to Qwen2.5 action format
eval/utils/qwen_mobile_tool.py:374
↓ 3 callersFunctionfind_best_resize
(original_size, scale_resolution, patch_size, allow_upscale=False)
sft/dataset.py:503
↓ 3 callersFunctionis_tap_action
(normalized_start_yx, normalized_end_yx)
eval/utils/evaluator.py:54
↓ 3 callersFunctionmake_context
( tokenizer: PreTrainedTokenizer, query: str, history: List[Tuple[str, str]] = None, syste
eval/utils/utils_odyssey/qwen_generation_utils.py:119
↓ 2 callersMethod__resize__
(self, origin_img)
sft/dataset.py:54
↓ 2 callersFunction_create_inputs
( processing_class, prompt_inputs, completions, )
rft/trainer/utils/process.py:48
↓ 2 callersMethod_enable_gradient_checkpointing
Enables gradient checkpointing for the model.
rft/trainer/arl.py:884
↓ 2 callersFunction_get_direction
(point1, point2)
eval/utils/evaluator.py:99
↓ 2 callersFunction_is_non_dual_point_action
(action_type)
eval/utils/action_utils.py:100
↓ 2 callersFunction_list_find
( input_list: List[Any], candidates: Tuple[Any], start: int = 0, )
eval/utils/utils_odyssey/tokenization_qwen.py:61
↓ 2 callersFunction_prepare_messages
( prompts, processing_class, max_prompt_length )
rft/trainer/utils/process.py:5
↓ 2 callersFunction_resize
(box: List[float])
eval/utils/evaluator.py:37
↓ 2 callersMethod_set_tcp_keepalive
设置TCP保持连接选项
rft/trainer/zmq.py:369
↓ 2 callersFunction_yx_in_bounding_boxes
Check if the (y,x) point is contained in each bounding box. Args: yx: The (y, x) coordinate in pixels of the point. bounding_boxe
eval/utils/action_utils.py:111
↓ 2 callersFunctionapply_rotary_pos_emb
(t, freqs)
eval/utils/utils_odyssey/modeling_qwen.py:1316
↓ 2 callersFunctionbuild_history_actions_str
(history_list)
eval/run_predict_ui_tars.py:77
↓ 2 callersFunctioncalculate_manhattan_distance
(x1, y1, x2, y2)
rft/trainer/utils/gui_eval.py:348
↓ 2 callersMethodcompute_loss
(self, model, inputs, return_outputs=False)
sft/trainer.py:14
↓ 2 callersFunctionconversation_to_ids
for single image multi-turn conversation conversation: [{'role': 'user', 'content': 'Describe this image'}, {'role': 'assi
sft/dataset.py:142
↓ 2 callersFunctiondecode_tokens
( tokens: Union[torch.LongTensor, TokensType], tokenizer: PreTrainedTokenizer, raw_text_len: in
eval/utils/utils_odyssey/qwen_generation_utils.py:265
↓ 2 callersFunctiondump_traj
(traj, out_root: Path, idx: int)
eval/eval_data/process_odyssey.py:120
↓ 2 callersFunctionextract_coords
(s)
eval/run_predict_ui_tars.py:298
↓ 2 callersFunctionget_1d_sincos_pos_embed_from_grid
embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)
eval/utils/utils_odyssey/modeling_qwen.py:161
↓ 2 callersFunctionget_1d_sincos_pos_embed_from_grid
embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)
eval/utils/utils_odyssey/visual.py:71
↓ 2 callersFunctionget_abs_pos
(abs_pos, tgt_size)
eval/utils/utils_odyssey/visual.py:23
↓ 2 callersFunctionget_direction
(point, to)
eval/utils/evaluator.py:138
↓ 2 callersFunctionget_stop_words_ids
(chat_format, tokenizer)
eval/utils/utils_odyssey/qwen_generation_utils.py:109
↓ 2 callersFunctionload
(path)
sft/finetune.py:103
↓ 2 callersFunctionload_json_data
(file_path)
eval/utils/convert_output.py:14
↓ 2 callersFunctionos_atlas_2minicpm
Convert a string containing low-level thinking and action information to minicpm schema format Args: action_str (str): String co
eval/run_predict_os_atlas.py:200
↓ 2 callersFunctionos_gensis_2minicpm
Convert a string containing low-level thoughts and action information into minicpm schema format Args: action_str (str): String cont
eval/run_predict_os_gensis.py:199
↓ 2 callersFunctionparse_action
(data)
eval/utils/convert_output.py:37
↓ 2 callersMethodprepare_generation
(self, model_wrapped, clear_device: bool = None, update_inference_model: bool = True)
rft/trainer/arl.py:519
↓ 2 callersFunctionqwen2_5_2_aitz
Convert Qwen2.5 output to AITZ output
eval/utils/qwen_mobile_tool.py:198
↓ 2 callersFunctionqwen2_5_2_minicpm
Convert Qwen2.5's output to minicpm's output
eval/run_predict_qwen2_5VL.py:302
↓ 2 callersMethodsample_step
(self, tasks, model: PreTrainedModel)
rft/trainer/arl.py:599
↓ 2 callersMethodsave_model
Will save the model, so you can reload it using `from_pretrained()`. Will only save from the main process.
rft/trainer/arl.py:930
↓ 2 callersMethodto_list_format
(self, text: str)
eval/utils/utils_odyssey/tokenization_qwen.py:376
↓ 2 callersFunctiontransform_action_data_and_build_test_data
(action_data)
eval/eval_data/process_ac.py:216
↓ 2 callersFunctionuitars2minicpm
Convert the ui-tars action string to the minicpm schema format Args: action_str (str): like "click(start_box='<|box_start|>(558,
eval/run_predict_ui_tars.py:285
↓ 1 callersMethod__init__
(self, img, scale=1.0)
eval/utils/utils_odyssey/tokenization_qwen.py:496
↓ 1 callersMethod__init__
(self, cfg=None)
eval/utils/utils_qwen/agent_function_call.py:84
↓ 1 callersMethod__init__
(self, jsonl_file_path: str, max_line_res: int|None = None, *args, **kwargs)
rft/trainer/utils/dataset.py:41
↓ 1 callersFunction__resize__
(origin_img)
eval/run_predict_minicpm.py:94
↓ 1 callersMethod_async_sampling
(self, unwrapped_model, epoch_iterator, num_batches)
rft/trainer/arl.py:383
↓ 1 callersMethod_attn
(self, query, key, value, registered_causal_mask, attention_mask=None, head_mask=None)
eval/utils/utils_odyssey/modeling_qwen.py:290
↓ 1 callersMethod_calc_stopped_samples
(self, prev_input_ids: Iterable[int])
eval/utils/utils_odyssey/qwen_generation_utils.py:374
↓ 1 callersFunction_check_drag_actions_match
Determines if two drag actions are the same.
eval/utils/action_utils.py:214
↓ 1 callersFunction_check_tap_actions_match
Determines if two tap actions are the same.
eval/utils/action_utils.py:183
↓ 1 callersMethod_click
(self, coordinate: Tuple[int, int])
eval/utils/utils_qwen/agent_function_call.py:122
↓ 1 callersFunction_decode_chatml
( tokens: List[int], *, stop_words: List[str], eod_token_ids: List[int], tokenizer: P
eval/utils/utils_odyssey/qwen_generation_utils.py:229
↓ 1 callersFunction_decode_default
( tokens: List[int], *, stop_words: List[str], eod_words: List[str], tokenizer: PreTr
eval/utils/utils_odyssey/qwen_generation_utils.py:196
↓ 1 callersMethod_disable_gradient_checkpointing
Disables gradient checkpointing for the model.
rft/trainer/arl.py:907
↓ 1 callersFunction_expand_mask
Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.
eval/utils/utils_odyssey/modeling_qwen.py:100
↓ 1 callersMethod_fetch_all_box_with_ref
(self, text)
eval/utils/utils_odyssey/tokenization_qwen.py:437
↓ 1 callersMethod_fetch_latest_picture
(self, response, history)
eval/utils/utils_odyssey/tokenization_qwen.py:427
↓ 1 callersMethod_get_local_indices
(total_size, world_size, rank)
eval/grounding_eval/code/Intern2.5-VL/evaluate_grounding_fun2bbox.py:100
↓ 1 callersMethod_get_local_indices
(total_size, world_size, rank)
eval/grounding_eval/code/Intern2.5-VL/evaluate_grounding_bbox2text.py:99
↓ 1 callersMethod_get_local_indices
(total_size, world_size, rank)
eval/grounding_eval/code/Intern2.5-VL/evaluate_grounding_text2bbox.py:95
↓ 1 callersMethod_get_train_sampler
Returns a sampler that ensures proper data sampling for GRPO training.
rft/trainer/arl.py:799
↓ 1 callersMethod_handle_ack
处理确认消息 (int)
rft/trainer/zmq.py:142
↓ 1 callersMethod_handle_queue_update
处理节点队列长度更新 (dict)
rft/trainer/zmq.py:169
↓ 1 callersMethod_handle_sync_request
处理同步优势请求 (SyncAdvantagesRequest)
rft/trainer/zmq.py:158
↓ 1 callersMethod_handle_task_status
处理任务状态 (TaskStatus)
rft/trainer/zmq.py:174
↓ 1 callersMethod_init_sockets
初始化所有ZMQ套接字和网络连接
rft/trainer/zmq.py:316
↓ 1 callersMethod_iter_data_sampling
(self, epoch_iterator, num_batches)
rft/trainer/arl.py:471
↓ 1 callersMethod_key
(self, text: str)
eval/utils/utils_qwen/agent_function_call.py:119
↓ 1 callersMethod_key
(self, keys: List[str])
eval/utils/utils_qwen/agent_function_call.py:252
↓ 1 callersMethod_left_click_drag
(self, coordinate: Tuple[int, int])
eval/utils/utils_qwen/agent_function_call.py:261
↓ 1 callersMethod_load_data_
(self)
eval/run_eval_agent.py:43
↓ 1 callersFunction_load_tiktoken_bpe
(tiktoken_bpe_file: str)
eval/utils/utils_odyssey/tokenization_qwen.py:53
↓ 1 callersMethod_long_press
(self, coordinate: Tuple[int, int], time: int)
eval/utils/utils_qwen/agent_function_call.py:125
↓ 1 callersFunction_make_causal_mask
Make causal mask used for bi-directional self-attention.
eval/utils/utils_odyssey/modeling_qwen.py:82
↓ 1 callersMethod_merge_heads
(self, tensor, num_heads, attn_head_size)
eval/utils/utils_odyssey/modeling_qwen.py:388
↓ 1 callersMethod_mouse_click
(self, button: str)
eval/utils/utils_qwen/agent_function_call.py:249
↓ 1 callersMethod_mouse_move
(self, coordinate: Tuple[int, int])
eval/utils/utils_qwen/agent_function_call.py:258
↓ 1 callersMethod_norm
(self, x)
eval/utils/utils_odyssey/modeling_qwen.py:1340
next →1–100 of 568, ranked by callers