Robertljg says the audit checked operations in PyTorch, JAX, TensorFlow and four other tools against TorchLean.
Softmax was only the opening act: Robertljg says an audit with @srush_nlp checked machine-learning operations across PyTorch, JAX, TensorFlow, NumPy, CuPy, ONNX Runtime and MLX against TorchLean. The post says they “found a lot more” beyond softmax.
1.3K
3 Sources, first seen ago
—
Not ranked yet
—
Not ranked yet
@Robertljgreally fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch, JAX, TF, NumPy, CuPy, ONNX Runtime and MLX, checked against TorchLean, and found a lot more!!!
@AnimaAnandkumarRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…
@patrickshaftoRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@Robertljgreally fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch, JAX, TF, NumPy, CuPy, ONNX Runtime and MLX, checked against TorchLean, and found a lot more!!!@AnimaAnandkumarRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@patrickshaftoRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@Robertljgreally fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch, JAX, TF, NumPy, CuPy, ONNX Runtime and MLX, checked against TorchLean, and found a lot more!!!@AnimaAnandkumarRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@patrickshaftoRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@Robertljgreally fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch, JAX, TF, NumPy, CuPy, ONNX Runtime and MLX, checked against TorchLean, and found a lot more!!!@AnimaAnandkumarRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…@patrickshaftoRT @Robertljg: really fun doing this with @srush_nlp!! softmax was just the start. we did a full numerical audit of ML ops across PyTorch,…
Comments