fix failed test cases for blt model#45596
Conversation
Signed-off-by: Liu, Kaixuan <kaixuan.liu@intel.com>
|
run-slow: blt |
|
This comment contains models: ["models/blt"] |
|
I update some expected values. @kaixuanliu You might want to run again with XPU, and update with |
| raise ValueError("input_ids is required for entropy-based patching") | ||
| _, patch_lengths, _ = self.patcher( | ||
| input_ids, | ||
| input_ids.to(self.patcher.embed_tokens.weight.device), |
There was a problem hiding this comment.
it's not clear why self.patcher doesn't handle the input device well.
There was a problem hiding this comment.
Likely the BltPatcher itself is BltPreTrainedModel
|
run-slow: blt |
|
This comment contains models: ["models/blt"] |
|
run-slow: blt |
|
This comment contains models: ["models/blt"] |
Signed-off-by: Liu, Kaixuan <kaixuan.liu@intel.com>
|
[For maintainers] Suggested jobs to run (before merge) run-slow: blt |
kaixuanliu
changed the title
|
Thx @ydshieh , I submitted another commit, and now all cases for blt can pass on XPU. |
2 participants
@ydshieh pls help review, thx!