hudabey commited on
Commit
bb6216a
·
verified ·
1 Parent(s): a39b100

Rename code

Browse files
Files changed (1) hide show
  1. README.md +15 -15
README.md CHANGED
@@ -1,7 +1,7 @@
1
  ---
2
  license: other
3
  license_name: apache-2.0-and-cc-by-nc-sa-4.0
4
- license_link: https://github.com/statite-ai/SpecAR#license
5
  pipeline_tag: text-to-video
6
  tags:
7
  - video-generation
@@ -32,9 +32,9 @@ This eliminates online perceptual routing and its associated scoring overhead, w
32
 
33
  StrideAR supports **text-to-video, image-to-video, and video-to-video generation** on Krea Realtime 14B, with a single implementation supporting NVIDIA Hopper and Blackwell GPUs.
34
 
35
- **[Project Page](https://statite.ai/stridear/) · [Earlier preprint](https://statite.ai/specar/specar.pdf) · [GitHub](https://github.com/statite-ai/SpecAR) · [ComfyUI](https://github.com/statite-ai/ComfyUI-SpecAR)**
36
 
37
- ![StrideAR speed comparison on Krea Realtime 14B](https://raw.githubusercontent.com/statite-ai/SpecAR/main/assets/race.webp)
38
 
39
  *Generation comparison on Krea Realtime 14B using one NVIDIA H100. The methods generate the same video duration from the same prompt and seed. Reported latencies come from the paper's evaluation.*
40
 
@@ -106,14 +106,14 @@ Requires Linux, Python 3.11, a compatible NVIDIA GPU and a CUDA 12.8-compatible
106
  ### Krea Realtime 14B
107
 
108
  ```bash
109
- git clone https://github.com/statite-ai/SpecAR
110
- cd SpecAR
111
 
112
  python3.11 -m venv .venv-krea
113
  source .venv-krea/bin/activate
114
 
115
  pip install ".[krea]"
116
- specar download --backbone krea
117
  ```
118
 
119
  The Krea setup downloads approximately 50 GB of public model files, including the Krea Realtime 14B target, Self-Forcing 1.3B drafter, text encoder and VAE.
@@ -129,7 +129,7 @@ On shared GPU machines, running `hf auth login` before downloading can help avoi
129
  Generate video directly from a text prompt.
130
 
131
  ```bash
132
- specar generate --backbone krea \
133
  --prompt "A papercraft coral reef full of colorful fish" \
134
  --k 2 \
135
  --out reef.mp4
@@ -142,7 +142,7 @@ By default, StrideAR generates 36 blocks, producing approximately 26.6 seconds o
142
  Generate a video conditioned on an input image.
143
 
144
  ```bash
145
- specar generate --backbone krea \
146
  --image photo.jpg \
147
  --prompt "A cinematic mountain landscape at sunrise" \
148
  --k 7 \
@@ -156,7 +156,7 @@ The image is resized to the model's 832×480 resolution and used as the initial
156
  Generate a video conditioned on an existing clip.
157
 
158
  ```bash
159
- specar generate --backbone krea \
160
  --video input.mp4 \
161
  --prompt "A futuristic city at night" \
162
  --strength 0.8 \
@@ -183,13 +183,13 @@ The `--strength` parameter controls the degree of transformation relative to the
183
 
184
  The default target budget is k=7. Lower values generally favor speed, while increasing the target budget changes the quality–latency trade-off.
185
 
186
- For all options and sample outputs, see the [StrideAR GitHub repository](https://github.com/statite-ai/SpecAR).
187
 
188
  ## ComfyUI
189
 
190
  StrideAR is also available as a ComfyUI custom-node integration for Krea Realtime 14B.
191
 
192
- **[ComfyUI-SpecAR — Installation, Nodes and Workflows](https://github.com/statite-ai/ComfyUI-SpecAR)**
193
 
194
  The ComfyUI integration uses the same StrideAR inference engine rather than a separate implementation of the decoding algorithm.
195
 
@@ -230,14 +230,14 @@ StrideAR is an approximate decoder: the scheduling guarantee does not guarantee
230
  The repository provides benchmark commands and reference latent hashes for reproducing the original H100 experiments.
231
 
232
  ```bash
233
- specar-benchmark --backbone krea --weights weights
234
  ```
235
 
236
  On the reference Hopper configuration with FlashAttention-3 and five denoising steps, the benchmark checks generated latents against the published SHA-256 hashes.
237
 
238
  Different attention kernels and GPU architectures can produce numerically different outputs. On other supported configurations, the benchmark records execution times and output hashes without requiring bitwise identity with the H100 reference.
239
 
240
- See the [repository documentation](https://github.com/statite-ai/SpecAR) for pinned dependencies, benchmark settings and Helios setup.
241
 
242
  ## Citation
243
 
@@ -259,10 +259,10 @@ The Krea Realtime backend incorporates code adapted from Krea's inference implem
259
 
260
  Model weights are not redistributed. Each checkpoint remains governed by its original license.
261
 
262
- See [LICENSE and file-level notices](https://github.com/statite-ai/SpecAR#license).
263
 
264
  ---
265
 
266
  Developed by **[Statite](https://statite.ai)**.
267
 
268
- [Project Page](https://statite.ai/stridear/) · [Earlier preprint](https://statite.ai/specar/specar.pdf) · [Source Code](https://github.com/statite-ai/SpecAR) · [ComfyUI](https://github.com/statite-ai/ComfyUI-SpecAR)
 
1
  ---
2
  license: other
3
  license_name: apache-2.0-and-cc-by-nc-sa-4.0
4
+ license_link: https://github.com/statite-ai/StrideAR#license
5
  pipeline_tag: text-to-video
6
  tags:
7
  - video-generation
 
32
 
33
  StrideAR supports **text-to-video, image-to-video, and video-to-video generation** on Krea Realtime 14B, with a single implementation supporting NVIDIA Hopper and Blackwell GPUs.
34
 
35
+ **[Project Page](https://statite.ai/stridear/) · [Paper](https://statite.ai/stridear/stridear.pdf) · [Earlier preprint](https://statite.ai/stridear/specar.pdf) · [GitHub](https://github.com/statite-ai/StrideAR) · [ComfyUI](https://github.com/statite-ai/ComfyUI-StrideAR)**
36
 
37
+ ![StrideAR speed comparison on Krea Realtime 14B](https://raw.githubusercontent.com/statite-ai/StrideAR/main/assets/race.webp)
38
 
39
  *Generation comparison on Krea Realtime 14B using one NVIDIA H100. The methods generate the same video duration from the same prompt and seed. Reported latencies come from the paper's evaluation.*
40
 
 
106
  ### Krea Realtime 14B
107
 
108
  ```bash
109
+ git clone https://github.com/statite-ai/StrideAR
110
+ cd StrideAR
111
 
112
  python3.11 -m venv .venv-krea
113
  source .venv-krea/bin/activate
114
 
115
  pip install ".[krea]"
116
+ stridear download --backbone krea
117
  ```
118
 
119
  The Krea setup downloads approximately 50 GB of public model files, including the Krea Realtime 14B target, Self-Forcing 1.3B drafter, text encoder and VAE.
 
129
  Generate video directly from a text prompt.
130
 
131
  ```bash
132
+ stridear generate --backbone krea \
133
  --prompt "A papercraft coral reef full of colorful fish" \
134
  --k 2 \
135
  --out reef.mp4
 
142
  Generate a video conditioned on an input image.
143
 
144
  ```bash
145
+ stridear generate --backbone krea \
146
  --image photo.jpg \
147
  --prompt "A cinematic mountain landscape at sunrise" \
148
  --k 7 \
 
156
  Generate a video conditioned on an existing clip.
157
 
158
  ```bash
159
+ stridear generate --backbone krea \
160
  --video input.mp4 \
161
  --prompt "A futuristic city at night" \
162
  --strength 0.8 \
 
183
 
184
  The default target budget is k=7. Lower values generally favor speed, while increasing the target budget changes the quality–latency trade-off.
185
 
186
+ For all options and sample outputs, see the [StrideAR GitHub repository](https://github.com/statite-ai/StrideAR).
187
 
188
  ## ComfyUI
189
 
190
  StrideAR is also available as a ComfyUI custom-node integration for Krea Realtime 14B.
191
 
192
+ **[ComfyUI-StrideAR — Installation, Nodes and Workflows](https://github.com/statite-ai/ComfyUI-StrideAR)**
193
 
194
  The ComfyUI integration uses the same StrideAR inference engine rather than a separate implementation of the decoding algorithm.
195
 
 
230
  The repository provides benchmark commands and reference latent hashes for reproducing the original H100 experiments.
231
 
232
  ```bash
233
+ stridear-benchmark --backbone krea --weights weights
234
  ```
235
 
236
  On the reference Hopper configuration with FlashAttention-3 and five denoising steps, the benchmark checks generated latents against the published SHA-256 hashes.
237
 
238
  Different attention kernels and GPU architectures can produce numerically different outputs. On other supported configurations, the benchmark records execution times and output hashes without requiring bitwise identity with the H100 reference.
239
 
240
+ See the [repository documentation](https://github.com/statite-ai/StrideAR) for pinned dependencies, benchmark settings and Helios setup.
241
 
242
  ## Citation
243
 
 
259
 
260
  Model weights are not redistributed. Each checkpoint remains governed by its original license.
261
 
262
+ See [LICENSE and file-level notices](https://github.com/statite-ai/StrideAR#license).
263
 
264
  ---
265
 
266
  Developed by **[Statite](https://statite.ai)**.
267
 
268
+ [Project Page](https://statite.ai/stridear/) · [Paper](https://statite.ai/stridear/stridear.pdf) · [Earlier preprint](https://statite.ai/stridear/specar.pdf) · [Source Code](https://github.com/statite-ai/StrideAR) · [ComfyUI](https://github.com/statite-ai/ComfyUI-StrideAR)