whn09 e10220bed0 Neuron: concatenate FP8 checkpoint tensors as bytes (fixes MoE on device) ... torch.cat of float8 tensors (e4m3 and e8m0) returns wrong data on the neuron device (torch-neuronx 2.14). The HF loader concatenates w1/w3 (and their UE8M0 scales) into the fused gate_up weights, so on the NeuronCore every routed and shared expert gate_up scale was zero and the MoE output vanished (issue #1). Concatenate 1-byte float dtypes through uint8 views, which is bit-identical on every platform. tests/neuron/device: create the neuron process group before the runtime initializes (torch-neuronx requirement) and add the float8 cat checks. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>		2026-10-21 11:46:58 +08:00
..
__init__.py	initial commit	2026-10-20 12:09:42 +08:00
config.py	initial commit	2026-10-20 12:09:42 +08:00
loading.py	initial commit	2026-10-20 12:09:42 +08:00
names.py	initial commit	2026-10-20 12:09:42 +08:00
parameters.py	initial commit	2026-10-20 12:09:42 +08:00
processor.py	initial commit	2026-10-20 12:09:42 +08:00
tokenizer.py	initial commit	2026-10-20 12:09:42 +08:00
weights.py	Neuron: concatenate FP8 checkpoint tensors as bytes (fixes MoE on device)	2026-10-21 11:46:58 +08:00