MCPcopy Create free account

hub / github.com/allenai/molmo2 / functions

Functions2,270 in github.com/allenai/molmo2

⨍ Functions 2,270◇ Types & classes 480
↓ 1 callersMethodget_html
(self)
olmo/eval/evaluators.py:95
↓ 1 callersFunctionget_image_files
(image_folder)
olmo/data/academic_video_track_datasets.py:109
↓ 1 callersMethodget_image_tokens
(self, image_grid: np.ndarray)
olmo/hf_model/processing_molmo2.py:114
↓ 1 callersFunctionget_key_subresponses
(response)
olmo/eval/mmmu_eval_utils.py:165
↓ 1 callersMethodget_legacy_key_mapping
Return a key map for legacy checkpoints
olmo/models/model.py:106
↓ 1 callersMethodget_level_text
(self, record: logging.LogRecord)
olmo/util.py:361
↓ 1 callersMethodget_location_text
(self, record: logging.LogRecord)
olmo/util.py:368
↓ 1 callersMethodget_max_grad_norm
( self, initial_max_grad_norm: Optional[float], step: int, max_steps: int )
olmo/train/optim.py:320
↓ 1 callersFunctionget_message
( images: list[Image.Image] | None, video_path: str | None, max_frames: int, frame_sample_mode
scripts/unified_demo.py:64
↓ 1 callersFunctionget_model
(checkpoint, model)
launch_scripts/sft.py:230
↓ 1 callersFunctionget_most_similar
Use the Levenshtein distance (or edit distance) to determine which of the choices is most similar to the given prediction
olmo/eval/math_vista_utils.py:45
↓ 1 callersMethodget_multi_choice_info
Given the list of options for multiple choice question Return the index2ans and all_choices https://github.com/MMMU-Benchmark
olmo/eval/evaluators.py:1911
↓ 1 callersFunctionget_node_rank
()
olmo/torch_util.py:63
↓ 1 callersMethodget_non_weight_decay_params
(self)
olmo/models/model.py:151
↓ 1 callersMethodget_output_shapes
(self)
olmo/models/molmo_point/molmo_point_example_preprocessor.py:542
↓ 1 callersMethodget_param_groups
Separate parameters into connector/vit/llm weight decay and non weight decay groups.
olmo/train/optim.py:106
↓ 1 callersFunctionget_parser
()
scripts/prepare_pretrained_model.py:576
↓ 1 callersFunctionget_recall_statements
( gt_statements: str, caption_str: str, openai_api_key: str )
olmo/eval/video_caption_utils.py:185
↓ 1 callersMethodget_replicate_and_shard_degree
(self, dp_world_size: int)
olmo/train/trainer_config.py:256
↓ 1 callersMethodget_rope_buffers_for_cp
Get the RoPE buffers to pass to each layer.
olmo/models/molmo2/molmo2.py:272
↓ 1 callersFunctionget_sampling_fps
Get the sampling fps that best spans the video and has the most frames sampled
olmo/data/video_loader.py:60
↓ 1 callersFunctionget_single_score
(result)
olmo/eval/tomato_eval.py:106
↓ 1 callersMethodget_system_prompt
(self, style, for_inference, messages, is_training, rng)
olmo/preprocessing/data_formatter.py:1537
↓ 1 callersMethodget_system_prompt
(self, style, for_inference, messages, is_training, rng)
olmo/models/molmo_point/molmo_point_data_formatter.py:729
↓ 1 callersFunctionget_target_fps
Get the target fps that best spans the video and has the most frames sampled
olmo/hf_model/video_processing_molmo2.py:498
↓ 1 callersMethodget_time_text
(self, record: logging.LogRecord)
olmo/util.py:356
↓ 1 callersFunctionget_training_mixture
(name)
launch_scripts/sft.py:287
↓ 1 callersMethodget_transparent_abc_image
(self, image, annotations)
olmo/data/dataset_builders/ai2d.py:121
↓ 1 callersFunctionget_video_hash
Get a hash of the video path for consistent sharding.
scripts/extract_ego4d_clips.py:40
↓ 1 callersMethodget_video_string
( self, video_grid: np.ndarray, timestamps: np.ndarray, )
olmo/hf_model/processing_molmo2.py:148
↓ 1 callersFunctiongpt_parser
(response, all_choices, index2ans)
olmo/eval/tomato_eval.py:14
↓ 1 callersMethodgumbel
Sample `Gumbel(phi)`. `phi` should have shape `(batch_size, num_classes)`.
olmo/nn/beam_search.py:524
↓ 1 callersFunctionhtml_rect
Utility method to get a HTML rectangle element
olmo/html_utils.py:545
↓ 1 callersFunctionimage_to_patches_and_grids
:return image_grids, the shape of each (low-res, high-res) image after pooling :return crops, the image crops to processes with the ViT :
olmo/hf_model/image_processing_molmo2.py:251
↓ 1 callersFunctionimage_to_patches_and_grids
:return image_grids, the shape of each image after pooling :return crops, the image crops to processes with the ViT :return pooled_patch_
olmo/hf_model/video_processing_molmo2.py:157
↓ 1 callersFunctioninit_process_group
()
olmo/torch_util.py:46
↓ 1 callersMethodinit_state
( self, start_class_log_probabilities: torch.Tensor, batch_size: int, num_classes: int )
olmo/nn/beam_search.py:434
↓ 1 callersMethodinit_state
( self, batch_size: int, )
olmo/nn/beam_search.py:850
↓ 1 callersFunctioninit_worker
()
olmo/data/dataset_builders/download_urls.py:62
↓ 1 callersMethodinitialize_and_load_model
(self)
olmo/eval/model_evaluator.py:274
↓ 1 callersMethodinsert_bos
Args: input_ids: [B, S] array with left padding attention_mask: [B, S] array (0 for pad, 1 for valid) bos
olmo/hf_model/processing_molmo2.py:181
↓ 1 callersFunctioninstall_excepthook
()
olmo/util.py:187
↓ 1 callersMethodis_point_in_mask
Check if a point is inside the mask.
olmo/eval/evaluators.py:2908
↓ 1 callersMethodlatest_checkpoint
Find the latest (sharded) checkpoint in a directory of checkpoints. :raises FileNotFoundError: If no checkpoints are found.
olmo/train/checkpointer.py:157
↓ 1 callersFunctionlist_sources
Print available video source names.
olmo/data/molmo2_video_track_datasets.py:2100
↓ 1 callersFunctionlistinstr
(lst, s, delimiter=None)
olmo/torch_util.py:215
↓ 1 callersFunctionllm_judge_score
(answers, pred, openai_api_key)
olmo/eval/llm_judge_utils.py:56
↓ 1 callersFunctionllm_output_to_rating
(llm_output: str)
olmo/eval/temp_compass_utils.py:72
↓ 1 callersMethodload
(self)
olmo/data/molmo2_video_track_datasets.py:521
↓ 1 callersMethodload
(self)
olmo/data/academic_image_datasets.py:948
↓ 1 callersFunctionload_and_shard_data
Load Ego4D clips metadata from JSONL file and return data for the specified shard. Args: clips_jsonl_path (str): Path to the JSO
scripts/extract_ego4d_clips.py:228
↓ 1 callersMethodload_mask
Load a binary mask from a PNG file.
olmo/data/academic_image_datasets.py:922
↓ 1 callersFunctionload_model_and_optim_state
Load model and optimizer state in-place from a checkpoint saved via :func:`save_model_and_optim_state()`. This method is agnostic to the dist
olmo/train/distributed_checkpointing.py:184
↓ 1 callersMethodload_trainer_state_dict
(self, state_dict: Dict[str, Any])
olmo/train/trainer.py:527
↓ 1 callersFunctionload_video
Loads `video` to a numpy array. Args: video (`VideoInput`): The video to convert to the numpy array format. Can be a lin
olmo/hf_model/video_processing_molmo2.py:435
↓ 1 callersFunctionload_video_torchcodec_dummy
( video_path: str, frame_sampler: TimeSampler, clip=None, seek_mode="exact" )
olmo/data/video_loader.py:447
↓ 1 callersMethodlog_batch
(self, batch)
olmo/train/trainer.py:160
↓ 1 callersMethodlog_progress
(self, on_step, target_step, eta=None)
olmo/train/trainer.py:139
↓ 1 callersFunctionmain
()
launch_scripts/eval_molmo2.py:126
↓ 1 callersFunctionmain
()
launch_scripts/eval.py:18
↓ 1 callersFunctionmain
()
launch_scripts/sft.py:435
↓ 1 callersFunctionmain
()
launch_scripts/train_gui_pointing.py:21
↓ 1 callersFunctionmain
(args: argparse.Namespace)
scripts/prepare_pretrained_model.py:559
↓ 1 callersFunctionmain
()
scripts/extract_epic_kitchens_clips.py:68
↓ 1 callersFunctionmain
()
scripts/extract_coin_clips.py:147
↓ 1 callersFunctionmain
()
scripts/dataset_visualize.py:67
↓ 1 callersFunctionmain
()
scripts/extract_ego4d_clips.py:318
↓ 1 callersFunctionmain
()
scripts/download_datasets.py:100
↓ 1 callersFunctionmain
()
olmo/hf_model/convert_molmo2_to_hf.py:481
↓ 1 callersFunctionmain
()
olmo/hf_model/test_molmo2.py:15
↓ 1 callersFunctionmain
()
olmo/hf_model/convert_molmo_point_to_hf.py:419
↓ 1 callersFunctionmain_llm
(args: argparse.Namespace)
scripts/prepare_pretrained_model.py:519
↓ 1 callersFunctionmain_vit
(args: argparse.Namespace)
scripts/prepare_pretrained_model.py:489
↓ 1 callersFunctionmantis_eval_mc
(target: str, prediction: str, options: List[str], question_type: str)
olmo/eval/vqa.py:713
↓ 1 callersFunctionmaybe_download_and_untar
(location, url, expected_dir=None)
olmo/data/utils.py:73
↓ 1 callersMethodmc_qoa_template
(data)
olmo/data/academic_video_datasets.py:3872
↓ 1 callersFunctionmean_relative_accuracy
( pred: float, target: float, start: float, end: float, interval: float, )
olmo/eval/vsi_bench_utils.py:25
↓ 1 callersFunctionmerge_overlapping_moments
Helper function to check overlap and merge moments (from Ego4d class).
scripts/extract_ego4d_clips.py:206
↓ 1 callersMethodmerge_visual_inputs
( self, input_ids: Optional[torch.LongTensor] = None, pixel_values: Optional[torch.Ten
olmo/hf_model/modeling_molmo_point.py:1324
↓ 1 callersMethodmerge_visual_inputs
( self, input_ids: Optional[torch.LongTensor] = None, pixel_values: Optional[torch.Ten
olmo/hf_model/modeling_molmo2.py:1396
↓ 1 callersFunctionmetaclip_resize
(image, desired_output_size)
olmo/preprocessing/image_preprocessor.py:204
↓ 1 callersFunctionmlvu_ssc_score
Calculate the accuracy and relevance scores for the SSC task in MLVU. Args: prediction (str): The response to evaluate. metadata (di
olmo/eval/mlvu_utils.py:104
↓ 1 callersFunctionmlvu_summary_score
Calculate the completeness and reliability scores for the Summary task in MLVU." Args: prediction (str): The response to evaluate. m
olmo/eval/mlvu_utils.py:140
↓ 1 callersFunctionmmmu_score
( target: List[str], response: str, metadata: dict, )
olmo/eval/vqa.py:303
↓ 1 callersMethodmodel_forward
( self, batch: Dict[str, Any], compute_z_loss: bool = False )
olmo/train/trainer.py:703
↓ 1 callersMethodmove_to_device
(self, batch, device)
olmo/train/trainer.py:677
↓ 1 callersFunctionmuir_bench_get_multi_choice_info
(options: List[str])
olmo/eval/vqa.py:594
↓ 1 callersFunctionmuir_bench_parse_multi_choice_response
Parse the prediction from the generated response. Return the predicted index e.g., A, B, C, D.
olmo/eval/vqa.py:607
↓ 1 callersFunctionnew_prune
Prune tokens based on their validity and similarity scores. Args: valid_mask (torch.Tensor): A mask indicating valid tokens. pool
olmo/nn/temporal_image_vit.py:743
↓ 1 callersFunctionnextqa_get_multi_choice_info
(options: List[str])
olmo/eval/vqa.py:508
↓ 1 callersFunctionnextqa_mc
(target: str, prediction: str, options: List[str])
olmo/eval/vqa.py:588
↓ 1 callersFunctionnextqa_parse_multi_choice_response
Parse the prediction from the generated response. Return the predicted index e.g., A, B, C, D. https://github.com/MMMU-Benchmark/MMMU/blo
olmo/eval/vqa.py:519
↓ 1 callersFunctionnormalize_image
( image: np.ndarray, image_mean: list[float], image_std: list[float], )
olmo/hf_model/video_processing_molmo2.py:51
↓ 1 callersFunctionnormalize_points
Convert absolute pixel coordinates to normalized coordinates. Args: points: Points to normalize (numpy array or list) sc
olmo/preprocessing/point_formatter.py:481
↓ 1 callersMethodone_file
Use the :data:`UnshardStrategy.one_file` strategy.
olmo/train/distributed_checkpointing.py:435
↓ 1 callersFunctionpack
(*examples: Dict)
olmo/data/dynamic_packer.py:168
↓ 1 callersMethodpad
( self, x: torch.Tensor, seq_dim: int, padding_to_add: int, value: Uni
olmo/nn/cp_load_balancer.py:453
↓ 1 callersFunctionparallelize_model
( model, *, world_mesh: Optional[DeviceMesh], compile_model: bool = False, float8_config =
olmo/dist_util.py:538
↓ 1 callersFunctionparse_caption_matching_output
Parse caption matching put based on word matching rules
olmo/eval/temp_compass_utils.py:201
↓ 1 callersFunctionparse_grounding_prediction_to_point_tracks
Parse point_ground_start_end model output. Expected format matches data_formatter.py: "0: ([34.0, 63.0, 0.50], [35.0, 64.0, 1.00
olmo/eval/object_tracking_utils.py:375
← previousnext →801–900 of 2,270, ranked by callers