[go: up one dir, main page]

0% found this document useful (0 votes)
53 views2 pages

Text To Image Vllms

Uploaded by

Faisal Aslam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
53 views2 pages

Text To Image Vllms

Uploaded by

Faisal Aslam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

| Model Name | Training Parameters (B) | Size (GB) | Accuracy (FID/%)

| Open Source | Custom Finetune |


|--------------------------|-------------------------|-----------|-----------------
-|-------------|-----------------|
| Imagen (Google) | 12B | 16 GB | ~8 FID
| No | No (API access) |
| MidJourney | Proprietary | Proprietary| ~8 FID
| No | No (Discord bot)|
| Muse (Google) | 1.2B | 1.5 GB | ~8 FID
| Yes | Yes |
| DeepFloyd IF | 1B | 5 GB | ~7.89 FID
| Yes | Yes |
| Parti (Google) | 20B | 40 GB | ~7.23 FID
| No | No |
| Stable Diffusion | 860M | 7 GB | ~10-15 FID
| Yes | Yes |
| Stable Diffusion 2.1 | 4B | 3.5 GB | ~10-15 FID
| Yes | Yes |
| DALL-E 2 | 12B | 16 GB | ~10 FID
| No | No (API access) |
| DALL-E | 12B | 20 GB | ~10.39 FID
| No | Yes |
| DALL-E 3 | ~12B | ~20 GB | Best-in-class
| No | No |
| GLIDE (OpenAI) | 3.5B | 10 GB | ~12.24 FID
| No | No |
| OpenJourney (Stable Diff.)| 0.98B | 4 GB | ~19 FID
| Yes | Yes |
| VQ-GAN+CLIP | Proprietary | Proprietary| ~15 FID
| Yes | Yes |
| Disco Diffusion | 1B | 3 GB | ~29.45 FID
| Yes | Yes |
| LAFITE (Tencent) | 1B | 2.5 GB | ~26.93 FID
| No | No |
| CICADA | 1B | 10 GB | 90.5% Accuracy
| Yes | Yes |
| VQ-VAE-2 | 1.2B | 15 GB | 89.5% Accuracy
| Yes | Yes |
| DALL-E 2 | 20B | 50 GB | 95.4% Accuracy
| No | Yes |
| Stable Diffusion | 860M | 7 GB | 93.5% Accuracy
| Yes | Yes |
| MidJourney | 1.2B | 30 GB | 92.5% Accuracy
| No | Yes |
| Imagen (Google) | 16B | 30 GB | 92.3% Accuracy
| No | No |
| DALL-E | 12B | 20 GB | 91.5% Accuracy
| No | Yes |
| DALL-E 3 | 350B | ~120 GB | 99.5% Accuracy
| No | No |
| MidJourney v5 | 250B | ~100 GB | 99.0% Accuracy
| No | No |
| Stable Diffusion XL | 6.6B | 6.5 GB | 98.5% Accuracy
| Yes | Yes |
| Imagen | 460B | ~150 GB | 98.0% Accuracy
| No | No |
| ERNIE-ViLG 2.0 | 10B | ~10 GB | 97.5% Accuracy
| No | No |
| Parti | 20B | ~20 GB | 97.0% Accuracy
| No | No |
| Stable Diffusion 2.1 | 1.5B | 5.3 GB | 96.5% Accuracy
| Yes | Yes |
| CogView2 | 10B | ~10 GB | 95.5% Accuracy
| Yes | Yes |
| Latent Diffusion | 1.4B | 4.3 GB | 95.0% Accuracy
| Yes | Yes |

You might also like