[model-free] raise error when ignoring quantized layer, allow special model (deepseekv4.1 engram.embed) #2631
feat: add validation for ignoring quantized NVFP4 layers and support …
9c93217a
feat: add indexed quantized source layer detection and corresponding …
08bdffcf
feat: enhance quantization handling with weight map loading and valid…
062034ae
Merge branch 'main' into xinhe/11-16a
9a0009a8
Merge branch 'main' into xinhe/11-16a
6ebe2572
fix: enhance ignored quantized source handling with model type support
bf96da8e
xin3he
changed the title feat: add validation for ignoring quantized NVFP4 layers and support for BF16 configuration [model-free] raise error when ignoring quantized layer, allow special model (deepseekv4.1 engram.embed) 1 day ago
xin3he
merged
1586bc6f
into main 1 day ago
xin3he
deleted the xinhe/11-16a branch 1 day ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub