Federico Cassano
Federico Cassano
It would be nice to have a way of getting these three in the z3 api. I know there are the bindings in z3-sys, but I would need to make...
### Description Hello, a simple binary I'm trying to patch segfaults. I believe this is due to some RIP relative addressing issue. The code I'm trying to patch in is...
Hello, it would be nice to get support for AsyncWrite and AsyncRead, either from Tokio or futures to serialize/deserialize rbxl binary and xml blobs. I would be happy to do...
We really like the streaming library. The only issue is that it leaves garbage around the file system after every run, and therefore we adopted this ugly hack: ```py MAX_RETRIES...
Hi! Wondering if there is any plan to support FP4 training in transformer engine. Would be great if so! Thanks!
Using FSDP2 with fp8_autocast seems to deadlock at the first forward pass through a `te.Linear`. By removing the autocast, we no longer get a deadlock. Using: - PyTorch version: `2.7.0.dev20250305+cu126`...
Hi, I ran my model with FSDP 2, one of the linear layers has a dim that's not divisible by the world size (128), and so I got the following...
Pipeline parallelism seem to hang when the number of microbatches is less than the degree. This issue occurs for both the standard and interleaved 1F1B schedules. Have not tested other...
When running the grouped gemm implementation and expert parallelism, i am faced with the following error: ``` [rank5]: File "/env/lib/python3.11/site-packages/megablocks-0.8.0.dev0-py3.11-linux-x86_64.egg/megablocks/layers/glu.py", line 255, in forward [rank5]: x1 = gg.ops.gmm(x, w1, batch_sizes,...
Some time ago I built this benchmark for MMTEB called CodeEditSearchRetrieval: https://github.com/embeddings-benchmark/mteb/blob/93a4db11371f1582c26277f12da84c1107296cd0/mteb/tasks/Retrieval/code/CodeEditSearchRetrieval.py#L26. It retrieves a relevant commit given an instruction. I found it quite useful for evaluating NL -> Code...