Open defertoexpertise opened 2 months ago
When i infer i get RuntimeError: Invalid buffer size: 50.62 GB
this is trying to use the FP32 model because FP16 wouldn't load... it gave the error "compute_index_ranges_weights" not implemented for 'Half'
Anyone know a workaround or how to get around this issue?
When i infer i get RuntimeError: Invalid buffer size: 50.62 GB
this is trying to use the FP32 model because FP16 wouldn't load... it gave the error "compute_index_ranges_weights" not implemented for 'Half'
Anyone know a workaround or how to get around this issue?