mirror of
https://github.com/tinygrad/tinygrad.git
synced 2026-06-24 02:14:17 +00:00
491e85597a* try to run commavq * fix 0 dim, start implementing new ops - Implement EmbedLayerNormalization - Implement Attention * SkipLayerNormalization and FastGelu * use original torch model, cast inputs * fix some ops: - properly do Cast - Attention: bi- and unidirectional - FastGelu: add bias before gelu * cleanup onnx_ops.py * add validation option to benchmark * cleanup imports * add checks incase onnx2torch implements ops in future * run onnx instead of original torch * just skip gpu on m1 * reactivate the other models * check for strange params & squash whitespace * cleanup * fix causal mask Attention * Range doesn't need int cast * embedding vocab_counter same dtype as input * no need to cast * always validate, fix PosixPath ort --------- Co-authored-by: George Hotz <george@comma.ai> |
||
|---|---|---|
| .. | ||
| accel | ||
| assembly | ||
| datasets | ||
| disk | ||
| dist | ||
| gemm | ||
| intel | ||
| ptx | ||
| rocm | ||
| triton | ||
| archprobe.py | ||
| augment.py | ||
| export_model.py | ||
| gradcheck.py | ||
| helpers.py | ||
| hip_wrapper.py | ||
| introspection.py | ||
| kernel_search.py | ||
| lib_test_ast.py | ||
| lr_scheduler.py | ||
| onnx.py | ||
| onnx_ops.py | ||
| thneed.py | ||
| training.py | ||
| utils.py | ||