Upload folder using huggingface_hub

flash_attn/impls/flash_attention.html

@@ -3847,6 +3847,7 @@ Cell: nv | 4.06s
  | <button class="run-btn" onclick="runCell('nv')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('nv')">Copy</button>
 <a href="cells/nv.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/flash_attention.md" target="_blank" class="github-btn">GitHub</a>
 </div>
 <div id="code-nv" class="cell-code" data-lines="3">
 <div class="code-wrap">
@@ -3908,6 +3909,7 @@ Cell: benchmark | 38.14s
  | <button class="run-btn" onclick="runCell('benchmark')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('benchmark')">Copy</button>
 <a href="cells/benchmark.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/flash_attention.md" target="_blank" class="github-btn">GitHub</a>
 </div>
 <div id="code-benchmark" class="cell-code" data-lines="66">
 <div class="code-wrap">

flash_attn/impls/hf_kernels_flash_attn3.html

@@ -3847,6 +3847,8 @@ Cell: benchmark | 40.68s
  | <button class="run-btn" onclick="runCell('benchmark')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('benchmark')">Copy</button>
 <a href="cells/benchmark.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/hf_kernels_flash_attn3.md" target="_blank" class="github-btn">GitHub</a>
+<a href="https://huggingface.co/kernels-community/flash-attn3" target="_blank" class="hf-btn">🤗 HF</a>
 </div>
 <div id="code-benchmark" class="cell-code" data-lines="71">
 <div class="code-wrap">

flash_attn/impls/mem_efficient_attention.html

@@ -3847,6 +3847,7 @@ Cell: benchmark | 39.23s
  | <button class="run-btn" onclick="runCell('benchmark')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('benchmark')">Copy</button>
 <a href="cells/benchmark.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/mem_efficient_attention.md" target="_blank" class="github-btn">GitHub</a>
 </div>
 <div id="code-benchmark" class="cell-code" data-lines="68">
 <div class="code-wrap">

flash_attn/impls/sage_attention.html

@@ -3847,6 +3847,7 @@ Cell: benchmark | 41.27s
  | <button class="run-btn" onclick="runCell('benchmark')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('benchmark')">Copy</button>
 <a href="cells/benchmark.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/sage_attention.md" target="_blank" class="github-btn">GitHub</a>
 </div>
 <div id="code-benchmark" class="cell-code" data-lines="79">
 <div class="code-wrap">

flash_attn/impls/xformers.html

@@ -3847,6 +3847,7 @@ Cell: benchmark | 41.87s
  | <button class="run-btn" onclick="runCell('benchmark')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('benchmark')">Copy</button>
 <a href="cells/benchmark.py" target="_blank" class="raw-btn">Raw</a>
+<a href="https://github.com/huggingface/kernels-uvnotes/blob/main/flash_attn/impls/xformers.md" target="_blank" class="github-btn">GitHub</a>
 </div>
 <div id="code-benchmark" class="cell-code" data-lines="68">
 <div class="code-wrap">

flash_attn/results/combined_results.html

@@ -4012,7 +4012,7 @@ using cross-file dependencies.</p>
   <rdf:RDF>
    <ns2:Work>
     <dc:type rdf:resource="http://purl.org/dc/dcmitype/StillImage" />
-    <dc:date>2025-10-14T20:33:46.842309</dc:date>
+    <dc:date>2025-10-14T20:47:15.973248</dc:date>
     <dc:format>image/svg+xml</dc:format>
     <dc:creator>
      <ns2:Agent>
@@ -5210,7 +5210,7 @@ using cross-file dependencies.</p>
 <span onclick="toggleOutput('combine')" style="cursor: pointer;">▶ output</span>
  <span id="uv-indicator-combine" onclick="toggleUvLogsFromHeader('combine')" style="cursor: pointer;">▶ uv-logs</span>
 </span> | 
-Cell: combine | 32.84s
+Cell: combine | 34.17s
  | <button class="run-btn" onclick="runCell('combine')">▶ run</button>
 <button class="copy-btn" onclick="copyCell('combine')">Copy</button>
 <a href="cells/combine.py" target="_blank" class="raw-btn">Raw</a>
@@ -5959,28 +5959,28 @@ Total records: 42
 <div class="uv-logs-header" onclick="toggleUvLogs(this)">▶ UV Install Logs</div>
 <div class="uv-logs-content" style="display: none;">
    Updating https://github.com/drbh/kernels-benchmark-tools.git (main)
-Downloading nvidia-cufile-cu12 (1.1MiB)
-Downloading nvidia-cufft-cu12 (184.2MiB)
-Downloading nvidia-cublas-cu12 (566.8MiB)
-Downloading networkx (1.9MiB)
+Downloading nvidia-curand-cu12 (60.7MiB)
 Downloading sympy (6.0MiB)
-Downloading nvidia-cuda-nvrtc-cu12 (84.0MiB)
+Downloading nvidia-cusolver-cu12 (255.1MiB)
 Downloading nvidia-cusparselt-cu12 (273.9MiB)
-Downloading nvidia-cudnn-cu12 (674.0MiB)
+Downloading nvidia-nccl-cu12 (307.4MiB)
+Downloading nvidia-cuda-nvrtc-cu12 (84.0MiB)
+Downloading nvidia-cufft-cu12 (184.2MiB)
+Downloading fonttools (4.7MiB)
+Downloading kiwisolver (1.4MiB)
+Downloading nvidia-cublas-cu12 (566.8MiB)
 Downloading nvidia-nvjitlink-cu12 (37.4MiB)
-Downloading numpy (15.9MiB)
-Downloading nvidia-cusolver-cu12 (255.1MiB)
-Downloading nvidia-curand-cu12 (60.7MiB)
-Downloading setuptools (1.1MiB)
 Downloading matplotlib (8.3MiB)
-Downloading fonttools (4.7MiB)
 Downloading nvidia-cusparse-cu12 (274.9MiB)
-Downloading pillow (6.3MiB)
-Downloading nvidia-nccl-cu12 (307.4MiB)
 Downloading nvidia-cuda-cupti-cu12 (9.8MiB)
-Downloading kiwisolver (1.4MiB)
-Downloading triton (148.4MiB)
+Downloading setuptools (1.1MiB)
+Downloading networkx (1.9MiB)
+Downloading numpy (15.9MiB)
+Downloading pillow (6.3MiB)
+Downloading nvidia-cufile-cu12 (1.1MiB)
 Downloading torch (846.8MiB)
+Downloading triton (148.4MiB)
+Downloading nvidia-cudnn-cu12 (674.0MiB)
     Updated https://github.com/drbh/kernels-benchmark-tools.git (f457279bca6573cd2fa54a74e67118f5e6b7a31c)
    Building kernels-benchmark-tools @ git+https://github.com/drbh/kernels-benchmark-tools.git@f457279bca6573cd2fa54a74e67118f5e6b7a31c
  Downloading nvidia-cufile-cu12
@@ -5993,20 +5993,20 @@ Downloading torch (846.8MiB)
  Downloading matplotlib
  Downloading nvidia-cuda-cupti-cu12
  Downloading numpy
- Downloading sympy
  Downloading nvidia-nvjitlink-cu12
+ Downloading sympy
  Downloading nvidia-curand-cu12
  Downloading nvidia-cuda-nvrtc-cu12
  Downloading triton
  Downloading nvidia-cufft-cu12
  Downloading nvidia-cusolver-cu12
- Downloading nvidia-cusparselt-cu12
  Downloading nvidia-cusparse-cu12
+ Downloading nvidia-cusparselt-cu12
  Downloading nvidia-nccl-cu12
  Downloading nvidia-cublas-cu12
  Downloading nvidia-cudnn-cu12
  Downloading torch
-Installed 37 packages in 254ms
+Installed 37 packages in 256ms
 </div>
 </div>
 <div class="cell-artifacts">
@@ -6800,7 +6800,7 @@ Installed 37 packages in 254ms
   <rdf:RDF>
    <ns2:Work>
     <dc:type rdf:resource="http://purl.org/dc/dcmitype/StillImage" />
-    <dc:date>2025-10-14T20:33:46.842309</dc:date>
+    <dc:date>2025-10-14T20:47:15.973248</dc:date>
     <dc:format>image/svg+xml</dc:format>
     <dc:creator>
      <ns2:Agent>