Functions — github.com/FoundationVision/LlamaGen

hub / github.com/FoundationVision/LlamaGen / functions

Functions465 in github.com/FoundationVision/LlamaGen

Copied from: https://github.com/numpy/numpy/blob/fb215c76967739268de71aa4bda55dd1b062bc2e/numpy/lib/format.py#L788-L886 Read from file-like

evaluations/c2i/evaluator.py:546

↓ 1 callersMethod_run_engine

(self, use_tqdm: bool)

autoregressive/serve/llm.py:238

↓ 1 callersFunction_sample

( probs: torch.Tensor, logprobs: torch.Tensor, sampling_metadata: SamplingMetadata, sampling_tensors:

autoregressive/serve/sampler.py:600

↓ 1 callersFunction_sample_with_torch

( probs: torch.Tensor, logprobs: torch.Tensor, sampling_metadata: SamplingMetadata, include_gp

autoregressive/serve/sampler.py:410

↓ 1 callersFunction_update_shapes

(pool3)

evaluations/c2i/evaluator.py:629

↓ 1 callersMethod_warm_up_model

(self)

autoregressive/serve/worker.py:189

↓ 1 callersMethodadd_request

Add a request to the engine's request pool. The request is added to the request pool and will be processed by the scheduler as `engin

autoregressive/serve/llm_engine.py:340

↓ 1 callersMethodbasic_clean

(text)

language/t5.py:91

↓ 1 callersFunctionbuild_coco

(args, transform)

dataset/coco.py:26

↓ 1 callersFunctionbuild_imagenet

(args, transform)

dataset/imagenet.py:53

↓ 1 callersFunctionbuild_imagenet_code

(args)

dataset/imagenet.py:56

↓ 1 callersFunctionbuild_openimage

(args, transform)

dataset/openimage.py:41

↓ 1 callersFunctionbuild_pexels

(args, transform)

dataset/pexels.py:3

↓ 1 callersFunctionbuild_t2i

(args, transform)

dataset/t2i.py:148

↓ 1 callersFunctionbuild_t2i_code

(args)

dataset/t2i.py:151

↓ 1 callersFunctionbuild_t2i_image

(args, transform)

dataset/t2i.py:145

↓ 1 callersMethodcache_swap

( self, blocks_to_swap_in: Dict[int, int], blocks_to_swap_out: Dict[int, int],

autoregressive/serve/worker.py:196

↓ 1 callersMethodcalculate_adaptive_weight

(self, nll_loss, g_loss, last_layer)

tokenizer/tokenizer_image/vq_loss.py:109

↓ 1 callersMethodcapture_model

Cuda graph capture a model. Note that CUDA graph's performance gain is negligible if number of batched tokens are larger than 200. An

autoregressive/serve/model_runner.py:982

↓ 1 callersFunctioncenter_crop_arr

Center cropping implementation from ADM. https://github.com/openai/guided-diffusion/blob/8fb3ad9197f16bbc40620447b2742e13458d2831/guided_diff

tokenizer/vae/reconstruction_vae_ddp.py:60

↓ 1 callersFunctioncenter_crop_arr

Center cropping implementation from ADM. https://github.com/openai/guided-diffusion/blob/8fb3ad9197f16bbc40620447b2742e13458d2831/guided_diff

tokenizer/vqgan/reconstruction_vqgan_ddp.py:62

↓ 1 callersFunctioncenter_crop_arr

Center cropping implementation from ADM. https://github.com/openai/guided-diffusion/blob/8fb3ad9197f16bbc40620447b2742e13458d2831/guided_diff

tokenizer/validation/val_ddp.py:55

↓ 1 callersFunctioncenter_crop_arr

Center cropping implementation from ADM. https://github.com/openai/guided-diffusion/blob/8fb3ad9197f16bbc40620447b2742e13458d2831/guided_diff

tokenizer/consistencydecoder/reconstruction_cd_ddp.py:60

↓ 1 callersFunctioncheck_image

(image_path)

tools/openimage_json.py:11

↓ 1 callersFunctioncompute_clip_score

( dataset: DataLoader, clip_model="ViT-B/32", device="cuda", how_many=5000)

evaluations/t2i/evaluation.py:130

↓ 1 callersFunctioncompute_entropy_loss

(affinity, loss_type="softmax", temperature=0.01)

tokenizer/tokenizer_image/vq_model.py:399

↓ 1 callersFunctioncompute_fid

(fake_dir: Path, gt_dir: Path, resize_size=None, feature_extractor="clip")

evaluations/t2i/evaluation.py:180

↓ 1 callersMethodcompute_inception_score

(self, activations: np.ndarray, split_size: int = 5000)

evaluations/c2i/evaluator.py:191

↓ 1 callersMethodcompute_logits

(self, hidden_states: torch.Tensor, sampling_metadata: SamplingMetadata)

autoregressive/serve/gpt_model.py:297

↓ 1 callersMethodcompute_prec_recall

( self, activations_ref: np.ndarray, activations_sample: np.ndarray )

evaluations/c2i/evaluator.py:206

↓ 1 callersMethodcompute_statistics

(self, activations: np.ndarray)

evaluations/c2i/evaluator.py:186

↓ 1 callersFunctioncreat_optimizer_by_name

(model, weight_decay, learning_rate, betas, global_rank, logger)

autoregressive/train/train_c2i_fsdp.py:67

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

tokenizer/vae/reconstruction_vae_ddp.py:41

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

tokenizer/vqgan/reconstruction_vqgan_ddp.py:43

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

tokenizer/validation/val_ddp.py:36

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

tokenizer/tokenizer_image/reconstruction_vq_ddp.py:25

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

tokenizer/consistencydecoder/reconstruction_cd_ddp.py:41

↓ 1 callersFunctioncreate_npz_from_sample_folder

Builds a single .npz file from a folder of .png samples.

autoregressive/sample/sample_c2i_ddp.py:21

↓ 1 callersMethodcustom_load_state_dict

(self, model_weights)

autoregressive/serve/gpt_model.py:311

↓ 1 callersFunctiondecode_n_tokens

( model, cur_token: torch.Tensor, input_pos: torch.Tensor, num_new_tokens: int, cfg_scale: float, cfg

autoregressive/models/generate.py:105

↓ 1 callersFunctiondecode_one_token

(model, x: torch.Tensor, input_pos: torch.Tensor, cfg_scale: float, cfg_flag: bool, **sampling_kwargs)

autoregressive/models/generate.py:89

↓ 1 callersMethoddetermine_num_available_blocks

Profiles the peak memory usage of the model to determine how many KV blocks may be allocated without OOMs. The engine will first cond

autoregressive/serve/worker.py:121

↓ 1 callersFunctiondownload

(url, local_path, chunk_size=1024)

tokenizer/tokenizer_image/lpips.py:24

↓ 1 callersFunctiondrop_path

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). This is the same as the DropConnect impl I created for E

utils/drop_path.py:4

↓ 1 callersMethodencode_request

( self, request_id: str, # pylint: disable=unused-argument prompt: Optional[str],

autoregressive/serve/llm_engine.py:326

↓ 1 callersFunctionevaluate_model

(opt)

evaluations/t2i/evaluation.py:207

↓ 1 callersMethodexecute_model

( self, seq_group_metadata_list: List[SequenceGroupMetadata], kv_caches: List[torch.Te

autoregressive/serve/model_runner.py:845

↓ 1 callersMethodexecute_model

( self, seq_group_metadata_list: Optional[List[SequenceGroupMetadata]] = None, blocks_

autoregressive/serve/worker.py:212

↓ 1 callersFunctionexists

(val)

tokenizer/tokenizer_image/discriminator.py:254

↓ 1 callersFunctionexists

(val)

tokenizer/tokenizer_image/discriminator_stylegan.py:100

↓ 1 callersFunctionfid_scaling_law_cfg

()

tools/draw_figure.py:43

↓ 1 callersFunctionfid_scaling_law_no_cfg

()

tools/draw_figure.py:6

↓ 1 callersFunctionfind_multiple

(n: int, k: int)

autoregressive/serve/gpt_model.py:18

↓ 1 callersMethodforward

( self, input_ids: torch.Tensor, positions: torch.Tensor, kv_caches: List[torc

autoregressive/serve/model_runner.py:1162

↓ 1 callersMethodfrom_engine_args

Creates an LLM engine from the engine arguments.

autoregressive/serve/llm_engine.py:255

↓ 1 callersMethodgen_fn

()

evaluations/c2i/evaluator.py:455

↓ 1 callersMethodget_cache_block_size_bytes

Get the size of the KV cache block size in bytes.

autoregressive/serve/worker.py:274

↓ 1 callersMethodget_codebook_entry

(self, indices, shape)

tokenizer/vqgan/quantize.py:92

↓ 1 callersMethodget_codebook_entry

(self, indices, shape=None, channel_first=True)

tokenizer/tokenizer_image/vq_model.py:261

↓ 1 callersMethodget_fsdp_wrap_module_list

(self)

autoregressive/models/gpt.py:385

↓ 1 callersMethodget_max_block_per_batch

(self)

autoregressive/serve/model_runner.py:244

↓ 1 callersMethodget_model_config

Gets the model configuration.

autoregressive/serve/llm_engine.py:458

↓ 1 callersMethodget_num_unfinished_requests

Gets the number of unfinished requests.

autoregressive/serve/llm_engine.py:462

↓ 1 callersMethodgetdata

(self, idx)

dataset/openimage.py:32

↓ 1 callersMethodgraph

(self)

autoregressive/serve/model_runner.py:1108

↓ 1 callersMethodhas_unfinished_requests

Returns True if there are unfinished requests.

autoregressive/serve/llm_engine.py:466

↓ 1 callersFunctioninit_worker_distributed_environment

Initialize the distributed environment.

autoregressive/serve/worker.py:282

↓ 1 callersMethodinitialize_cache

Allocate GPU and CPU KV cache with the specified number of blocks. This also warms up the model, which may record CUDA graphs.

autoregressive/serve/worker.py:166

↓ 1 callersMethodinitialize_weights

(self)

autoregressive/models/gpt.py:300

↓ 1 callersMethodless_thans

(self, batch_1, radii_1, batch_2, radii_2)

evaluations/c2i/evaluator.py:414

↓ 1 callersMethodload_dataset

(self)

evaluations/t2i/evaluation.py:82

↓ 1 callersMethodload_from_pretrained

(self, name="vgg_lpips")

tokenizer/tokenizer_image/lpips.py:69

↓ 1 callersFunctionload_model

(args)

app.py:25

↓ 1 callersFunctionload_model

(args)

tools/push_vae_to_hf.py:17

↓ 1 callersMethodload_model

(self, args)

autoregressive/serve/worker.py:117

↓ 1 callersFunctionmain

(args)

tools/push_gpt_to_hf.py:13

↓ 1 callersFunctionmain

(args)

tools/check_image_codes.py:9

↓ 1 callersFunctionmain

(args)

tools/openimage_json.py:44

↓ 1 callersFunctionmain

()

evaluations/c2i/evaluator.py:27

↓ 1 callersFunctionmain

Trains a new DiT model.

language/extract_t5_feature.py:53

↓ 1 callersFunctionmain

(args)

tokenizer/vae/sd_vae_demo.py:9

↓ 1 callersFunctionmain

(args)

tokenizer/vae/reconstruction_vae_ddp.py:81

↓ 1 callersFunctionmain

(args)

tokenizer/vqgan/reconstruction_vqgan_ddp.py:83

↓ 1 callersFunctionmain

(args)

tokenizer/vqgan/taming_vqgan_demo.py:17

↓ 1 callersFunctionmain

(args)

tokenizer/validation/val_ddp.py:76

↓ 1 callersFunctionmain

(args)

tokenizer/tokenizer_image/reconstruction_vq_ddp.py:43

↓ 1 callersFunctionmain

(args)

tokenizer/tokenizer_image/vq_demo.py:13

↓ 1 callersFunctionmain

Trains a new model.

tokenizer/tokenizer_image/vq_train.py:36

↓ 1 callersFunctionmain

(args)

tokenizer/consistencydecoder/reconstruction_cd_ddp.py:81

↓ 1 callersFunctionmain

(args)

tokenizer/consistencydecoder/cd_demo.py:9

↓ 1 callersFunctionmain

(args)

autoregressive/serve/sample_c2i.py:12

↓ 1 callersFunctionmain

(args)

autoregressive/sample/sample_t2i.py:20

↓ 1 callersFunctionmain

(args)

autoregressive/sample/sample_c2i_ddp.py:38

↓ 1 callersFunctionmain

(args)

autoregressive/sample/sample_c2i.py:18

↓ 1 callersFunctionmain

(args)

autoregressive/sample/sample_t2i_ddp.py:26

↓ 1 callersFunctionmain

(args)

autoregressive/train/extract_codes_c2i.py:23

↓ 1 callersFunctionmain

(args)

autoregressive/train/train_c2i_fsdp.py:102

↓ 1 callersFunctionmain

Trains a new DiT model.

autoregressive/train/extract_codes_t2i.py:56

↓ 1 callersFunctionmain

(args)

autoregressive/train/train_c2i.py:57

↓ 1 callersFunctionmain

(args)

autoregressive/train/train_t2i.py:27

← previous next →101–200 of 465, ranked by callers