dingshizhe/vgtae-minit2i-sft-latents
MiniT2I SFT-set latents encoded with VGT-AE (448px) The text-to-image SFT mixture, pre-encoded into the VGT-AE latent space. Companion to dingshizhe/vgtae-cc12m-latents, which covers the pretraining corpus; same encoder, same settings, same layout. These are not DC-AE latents. VGT-AE is a hybrid codec — a fine-tuned Qwen2.5-VL ViT as the encoder, a DC-AE decoder at sampling time. The tensor shape coincides with DC-AE f32c32 (32, 16, 16), but the space is different.… See the full description on the dataset page: https://huggingface.co/datasets/dingshizhe/vgtae-minit2i-sft-latents.
MiniT2I SFT-set latents encoded with VGT-AE (448px)
The text-to-image SFT mixture, pre-encoded into the VGT-AE latent space. Companion to `dingshizhe/vgtae-cc12m-latents`, which covers the pretraining corpus; same encoder, same settings, same layout.
These are not DC-AE latents. VGT-AE is a hybrid codec — a fine-tuned Qwen2.5-VL ViT as the encoder, a DC-AE decoder at sampling time. The tensor shape coincides with DC-AE f32c32 (32, 16, 16), but the space is different.
Contents
Two subsets, each a set of WebDataset tars aligned shard-for-shard with the source:
blip3_ft60k/journeyDB.tar
<key>.npy float16, shape (32, 16, 16)
<key>.txt captionblip3_ft60k is itself a mixture — its per-source shards are dalle3, geneval_train, human_gestures, journeyDB, mscoco_human, object_1/2, occupation_1/2, text_1/2. Note that the shard named dalle3.tar inside blip3_ft60k/ is not the same data as the top-level dalle3/ subset.
Encoder settings
Integrity
Every output tar was walked end to end after encoding: 0 unreadable or truncated shards, and the per-shard .npy counts sum to 77,883 — exactly the number of images in the source tars, so nothing was dropped.
Usage
import io, tarfile, numpy as np
with tarfile.open("blip3_ft60k/journeyDB.tar") as tar:
for m in tar:
if m.name.endswith(".npy"):
z = np.load(io.BytesIO(tar.extractfile(m).read())) # (32, 16, 16) fp16