Oct 21 Yesterday 18:33 Model card: note torch.compile comparison Adds one sentence under the results table: torch.compile vs eager (1.6–2.6×) and GEAK vs torch.compile for FastWAM (3.7×), as reported in the ROCm blog. Updates the README.md line in ASSET_SHA256SUMS. 18:27 Model card: link the ROCm blog, label success rate (#2 ) - Model card: link the ROCm blog, label success rate (a600ee39ed0b9cae9cc8b7ba53604b7238efc956) 18:22 Model card: results, download and launcher variables (#1 ) - Model card: results, download and launcher variables (605cea68c4bf0d06aebacb0964a0ce25429f1f48) Oct 15 7 days ago 20:51 Describe the Compact Action Expert and its intended use. 17:24 Add the CAE-FastWAM citation. 17:19 Add replay.mp4 for the model page video preview. 17:16 Show the LIBERO videos in the model card gallery. 17:10 Add LIBERO example videos. 13:00 Link the AMD-AGI faster-action-models repository. 12:50 Remove the reported success rate from the model card. 12:32 Rename the model card to CAE-FastWAM. 12:23 State that no training data was added beyond the FastWAM recipe. 12:10 Restore the statement that the Wan2.2 weights are fine-tuned. 12:08 State that the Wan2.2 weights are used unchanged. 12:06 Update the model card checksum. 12:06 State that the FastWAM LIBERO training data is used unchanged. Oct 13 9 days ago 11:05 Add files using upload-large-folder tool 11:04 initial commit 
