Skip to content

Commit a3dafe4

Browse files
Amar3ttoAmar3tto
committed
Rename files
1 parent 006f650 commit a3dafe4

6 files changed

Lines changed: 17 additions & 17 deletions

File tree

‎.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Image_Classification_Rightfit.txt‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -42,5 +42,5 @@
4242
--inference_batch_size=auto
4343
--window_sec=60
4444
--trigger_proc_time_sec=30
45-
--enable_dedup=false
45+
--enable_dedup=true
4646
--experiments=worker_accelerator=type:nvidia-tesla-t4;count:1;install-nvidia-driver:5xx

sdks/python/apache_beam/ml/inference/pytorch_image_captioning_blip_clip_requirements.txt renamed to sdks/python/apache_beam/ml/inference/pytorch_image_captioning_requirements.txt

File renamed without changes.

‎website/www/site/content/en/performance/_index.md‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -58,5 +58,5 @@ See the following pages for performance measures recorded when running various B
5858
- [PyTorch Vision Classification Resnet 152 Tesla T4 GPU](/performance/pytorchresnet152tesla)
5959
- [TensorFlow MNIST Image Classification](/performance/tensorflowmnist)
6060
- [VLLM Gemma Batch Completion Tesla T4 GPU](/performance/vllmgemmabatchtesla)
61-
- [PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch](/performance/pytorchimageobjectdetection)
62-
- [PyTorch Image Captioning BLIP + CLIP Batch](/performance/pytorchimagecaptioning)
61+
- [PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch GPU](/performance/pytorchimageobjectdetectionbatchgpu)
62+
- [PyTorch Image Captioning BLIP + CLIP Batch GPU](/performance/pytorchimagecaptioningbatchgpu)

website/www/site/content/en/performance/pytorchimagecaptioning/_index.md renamed to website/www/site/content/en/performance/pytorchimagecaptioningbatchgpu/_index.md

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
---
2-
title: "PyTorch Image Captioning BLIP + CLIP Batch Performance"
2+
title: "PyTorch Image Captioning BLIP + CLIP Batch GPU Performance"
33
---
44

55
<!--
@@ -14,7 +14,7 @@ See the License for the specific language governing permissions and
1414
limitations under the License.
1515
-->
1616

17-
# PyTorch Image Captioning BLIP + CLIP Batch
17+
# PyTorch Image Captioning BLIP + CLIP Batch GPU
1818

1919
**Model**: PyTorch Image Captioning — BLIP (candidate generation) + CLIP (ranking)
2020
**Accelerator**: Tesla T4 GPU (fixed batch size)
@@ -24,19 +24,19 @@ This batch pipeline performs image captioning using a multi-model open-source Py
2424
It first generates multiple candidate captions per image using a BLIP model, then ranks these candidates with a CLIP model based on image-text similarity.
2525
The pipeline runs batched GPU inference with fixed batch sizes and ensures exactly-once semantics through deterministic input deduplication and file-based BigQuery writes, enabling stable and reproducible performance measurements across batch runs.
2626

27-
The following graphs show various metrics when running PyTorch Image Captioning BLIP + CLIP Batch pipeline.
27+
The following graphs show various metrics when running PyTorch Image Captioning BLIP + CLIP Batch GPU pipeline.
2828
See the [glossary](/performance/glossary) for definitions.
2929

3030
Full pipeline implementation is available [here](https://github.com/apache/beam/blob/master/sdks/python/apache_beam/examples/inference/pytorch_image_captioning.py).
3131

3232
## What is the estimated cost to run the pipeline?
3333

34-
{{< performance_looks io="pytorchimagecaptioning" read_or_write="write" section="cost" >}}
34+
{{< performance_looks io="pytorchimagecaptioningbatchgpu" read_or_write="write" section="cost" >}}
3535

3636
## How has various metrics changed when running the pipeline for different Beam SDK versions?
3737

38-
{{< performance_looks io="pytorchimagecaptioning" read_or_write="write" section="version" >}}
38+
{{< performance_looks io="pytorchimagecaptioningbatchgpu" read_or_write="write" section="version" >}}
3939

4040
## How has various metrics changed over time when running the pipeline?
4141

42-
{{< performance_looks io="pytorchimagecaptioning" read_or_write="write" section="date" >}}
42+
{{< performance_looks io="pytorchimagecaptioningbatchgpu" read_or_write="write" section="date" >}}

website/www/site/content/en/performance/pytorchimageobjectdetection/_index.md renamed to website/www/site/content/en/performance/pytorchimageobjectdetectionbatchgpu/_index.md

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
---
2-
title: "PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch Performance"
2+
title: "PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch GPU Performance"
33
---
44

55
<!--
@@ -14,7 +14,7 @@ See the License for the specific language governing permissions and
1414
limitations under the License.
1515
-->
1616

17-
# PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch
17+
# PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch GPU
1818

1919
**Model**: PyTorch Image Object Detection — Faster R-CNN ResNet-50 FPN (pretrained on COCO)
2020
**Accelerator**: Tesla T4 GPU (fixed batch size)
@@ -24,19 +24,19 @@ This batch pipeline performs object detection using an open-source PyTorch Faste
2424
It reads image URIs from GCS, decodes and preprocesses images, and runs batched inference with a fixed batch size to measure stable GPU performance.
2525
The pipeline ensures exactly-once semantics within batch execution by deduplicating inputs and writing results to BigQuery using file-based loads, enabling reproducible and comparable performance measurements across runs.
2626

27-
The following graphs show various metrics when running PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch pipeline.
27+
The following graphs show various metrics when running PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch GPU pipeline.
2828
See the [glossary](/performance/glossary) for definitions.
2929

3030
Full pipeline implementation is available [here](https://github.com/apache/beam/blob/master/sdks/python/apache_beam/examples/inference/pytorch_image_object_detection.py).
3131

3232
## What is the estimated cost to run the pipeline?
3333

34-
{{< performance_looks io="pytorchimageobjectdetection" read_or_write="write" section="cost" >}}
34+
{{< performance_looks io="pytorchimageobjectdetectionbatchgpu" read_or_write="write" section="cost" >}}
3535

3636
## How has various metrics changed when running the pipeline for different Beam SDK versions?
3737

38-
{{< performance_looks io="pytorchimageobjectdetection" read_or_write="write" section="version" >}}
38+
{{< performance_looks io="pytorchimageobjectdetectionbatchgpu" read_or_write="write" section="version" >}}
3939

4040
## How has various metrics changed over time when running the pipeline?
4141

42-
{{< performance_looks io="pytorchimageobjectdetection" read_or_write="write" section="date" >}}
42+
{{< performance_looks io="pytorchimageobjectdetectionbatchgpu" read_or_write="write" section="date" >}}

‎website/www/site/data/performance.yaml‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -266,7 +266,7 @@ looks:
266266
title: AvgThroughputBytesPerSec by Version
267267
- id: VJMWrZh3jXk2mqCZk4NQn3tBrHgGWqnC
268268
title: AvgThroughputElementsPerSec by Version
269-
pytorchimageobjectdetection:
269+
pytorchimageobjectdetectionbatchgpu:
270270
write:
271271
folder: #TODO
272272
cost:
@@ -282,7 +282,7 @@ looks:
282282
title: AvgThroughputBytesPerSec by Version
283283
- id: #TODO
284284
title: AvgThroughputElementsPerSec by Version
285-
pytorchimagecaptioning:
285+
pytorchimagecaptioningbatchgpu:
286286
write:
287287
folder: #TODO
288288
cost:

0 commit comments

Comments
 (0)