Downloads · 30 days
0
ldlm-flowmap-semi
elephantmipt/ldlm-flowmap-semi
ldlm-flowmap-semi is a text generation model from elephantmipt. Use it when you need the model to write or continue text. It is set up for pytorch.
GPT-2 latent diffusion language-model checkpoints trained on OpenWebText.
Access
Public
Updated Sep 24, 2026
Repo size
66.5 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pth66.5 GB · 100%
At a glance
- Task
- Text Generation
- Library
- pytorch
- Access
- Public
- Created
- Sep 8, 2026
- Updated
- Sep 24, 2026
- SHA
- b4adb484
Try a prompt
Model card
From the Hugging Face model README
LDLM flow-map semi-supervised checkpoints
GPT-2 latent diffusion language-model checkpoints trained on OpenWebText.
N=8 adaptive-SDE evaluation
All values use eight direct flow-map hops on the adaptive time grid with the
gamma SDE rule (eta=1). Lower perplexity is better. The 50k columns show
early-training behavior; the final columns show metrics at step 500k.
| Links | Variant | PPL @ 50k | Entropy @ 50k | PPL @ final | Entropy @ final |
|---|---|---|---|---|---|
| link_final | Diagonal-only | 14.30 | 1.4031 | 231.00 | 2.9367 |
| link_final | Frozen AE, flow map initialized from diagonal-only 150k | 54.33 | 4.1346 | 60.23 | 4.1387 |
| link_final | Unfrozen AE, flow map initialized from diagonal-only 150k | 110.33 | 4.2566 | 69.74 | 4.1420 |
| link_final link_50k | Frozen AE, initialized from its previous final checkpoint | 59.97 | 4.1163 | 60.78 | 4.1523 |
| link_final link_50k | Unfrozen AE, initialized from its previous final checkpoint | 92.58 | 4.2229 | 80.64 | 4.1702 |
| link_final link_50k | Frozen AE, final initialization, diagonal-only teacher target | 165.67 | 4.3389 | 361.59 | 4.1609 |
| link_final link_50k | Unfrozen AE, final initialization, diagonal-only teacher target | 849.46 | 2.6878 | 716.32 | 3.0548 |
The teacher-target checkpoints use the EMA model from the diagonal-only 500k run as the detached diagonal target. The target is represented as the teacher's predicted clean latent, which is the x0 parameterization used by the flow-map sampler for its instantaneous probability-flow velocity.
Initialization artifact: diagonal-only 150k snapshot.
Experiment logs
- Diagonal-only
- Frozen AE, flow map initialized from diagonal-only 150k
- Unfrozen AE, flow map initialized from diagonal-only 150k
- Frozen AE, initialized from its previous final checkpoint
- Unfrozen AE, initialized from its previous final checkpoint
- Frozen AE, final initialization, diagonal-only teacher target
- Unfrozen AE, final initialization, diagonal-only teacher target
- Task
- Text Generation
- Library
- pytorch
- Created
- Sep 8, 2026
- Updated
- Sep 24, 2026