shahafvl commited on
Commit
edec36a
·
verified ·
1 Parent(s): ca9a164

README: add prominent GH link + fix stale Objaverse_Splats URL in Lineage

Browse files
Files changed (1) hide show
  1. README.md +3 -1
README.md CHANGED
@@ -30,13 +30,15 @@ decoder) is otherwise unchanged.
30
  This checkpoint is **V10b epoch 26**: an LPIPS-VGG perceptual fine-tune (loss
31
  weight 0.2) of V9 epoch 60.
32
 
 
 
33
  ## Lineage
34
 
35
  - Trained from the RenderFormer architecture (`microsoft/renderformer-v1.1-swin-large`)
36
  with the mesh input encoder replaced by a learned linear projection over
37
  14-d Gaussian parameters: `[pos(3), scale(3), quat(4), rgb(3), opacity(1)]`.
38
  - **V9 (pretrain).** Single-object 3DGS scans from
39
- [Objaverse_Splats](https://huggingface.co/datasets/zerogu/Objaverse_Splats)
40
  (2,667 train / 183 val), pure log-HDR L1 loss, 60 epochs on 3 GPUs.
41
  - **V10b (this checkpoint).** Fine-tuned from V9 ep60 with a combined log-HDR
42
  L1 + LPIPS-VGG loss on tonemapped LDR output, 26 epochs on 4 GPUs (cosine LR
 
30
  This checkpoint is **V10b epoch 26**: an LPIPS-VGG perceptual fine-tune (loss
31
  weight 0.2) of V9 epoch 60.
32
 
33
+ **Code & docs:** [github.com/SVLwoof/gaussianformer](https://github.com/SVLwoof/gaussianformer)
34
+
35
  ## Lineage
36
 
37
  - Trained from the RenderFormer architecture (`microsoft/renderformer-v1.1-swin-large`)
38
  with the mesh input encoder replaced by a learned linear projection over
39
  14-d Gaussian parameters: `[pos(3), scale(3), quat(4), rgb(3), opacity(1)]`.
40
  - **V9 (pretrain).** Single-object 3DGS scans from
41
+ [Objaverse_Splats](https://huggingface.co/datasets/ShapeSplats/Objaverse_Splats)
42
  (2,667 train / 183 val), pure log-HDR L1 loss, 60 epochs on 3 GPUs.
43
  - **V10b (this checkpoint).** Fine-tuned from V9 ep60 with a combined log-HDR
44
  L1 + LPIPS-VGG loss on tonemapped LDR output, 26 epochs on 4 GPUs (cosine LR