SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2308.06571 · 2023

ModelScope Text-to-Video Technical Report

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 16 functions out of this paper's own repositories and ran 11 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
exponentialml/text-to-video-finetuning canonical 11 of 16
FunctionStatusWhere it lives
convert_text_enc_state_dict Ran exponentialml/text-to-video-finetuning/utils/convert_diffusers_to_original_ms_text_to_video.py
code served (permissive licence) · get_code("a59539766c1c5a5a")
convert_text_enc_state_dict_v20 Ran exponentialml/text-to-video-finetuning/utils/convert_diffusers_to_original_ms_text_to_video.py
code served (permissive licence) · get_code("8d4b0c9f8b9a884e")
download_progress Ran exponentialml/text-to-video-finetuning/utils/lama.py
code served (permissive licence) · get_code("bfdb56f5b6ab8034")
filter_dict Ran exponentialml/text-to-video-finetuning/utils/lora_handler.py
code served (permissive licence) · get_code("9191602362628ded")
get_bucket_sizes Ran exponentialml/text-to-video-finetuning/utils/bucketing.py
code served (permissive licence) · get_code("8ff113de21fb6bc2")
get_prompt_ids Ran exponentialml/text-to-video-finetuning/utils/dataset.py
code served (permissive licence) · get_code("6f7c993202e44dda")
is_video Ran exponentialml/text-to-video-finetuning/models/unet_3d_blocks.py
code served (permissive licence) · get_code("c1232f9d1e7794e4")
min_res Ran exponentialml/text-to-video-finetuning/utils/bucketing.py
code served (permissive licence) · get_code("40775842d99a25f9")
normalize_input Ran exponentialml/text-to-video-finetuning/utils/dataset.py
code served (permissive licence) · get_code("dbfc36f5a5a5a1fd")
read_caption_file Ran exponentialml/text-to-video-finetuning/utils/dataset.py
code served (permissive licence) · get_code("00090fcbdd45d352")
up_down_bucket Ran exponentialml/text-to-video-finetuning/utils/bucketing.py
code served (permissive licence) · get_code("bec7ef14686681c0")
convert_unet_state_dict Not yet run exponentialml/text-to-video-finetuning/utils/convert_diffusers_to_original_ms_text_to_video.py
code served (permissive licence) · get_code("754cea3454b8b8da")
inject_inferable_lora Not yet run exponentialml/text-to-video-finetuning/utils/lora.py
code served (permissive licence) · get_code("df1f737e7e9ec97f")
inject_trainable_lora Not yet run exponentialml/text-to-video-finetuning/utils/lora.py
code served (permissive licence) · get_code("da50048edb81475c")
inject_trainable_lora_extended Not yet run exponentialml/text-to-video-finetuning/utils/lora.py
code served (permissive licence) · get_code("929d4d45fed75939")
inpaint_watermark Not yet run exponentialml/text-to-video-finetuning/utils/lama.py
code served (permissive licence) · get_code("3c2dce662245aae1")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

This paper introduces ModelScopeT2V, a text-to-video synthesis model that evolves from a text-to-image synthesis model (i.e., Stable Diffusion). ModelScopeT2V incorporates spatio-temporal blocks to ensure consistent frame generation and smooth movement transitions. The model could adapt to varying frame numbers during training and inference, rendering it suitable for both image-text and video-text datasets. ModelScopeT2V brings together three components (i.e., VQGAN, a text encoder, and a denoising UNet), totally comprising 1.7 billion parameters, in which 0.5 billion parameters are dedicated to temporal capabilities. The model demonstrates superior performance over state-of-the-art methods across three evaluation metrics. The code and an online demo are available at \url{https://modelscope.cn/models/damo/text-to-video-synthesis/summary}.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2308.06571")
get_code_for_paper("2308.06571")
have("2308.06571")

Connect an agent — have() is free.