From: Michał Piszczek <redacted>
Date: Fri, 8 May 2026 03:55:48 +0000 (+0200)
Subject: convert : fix RuntimeError when stripping FP8 KV-cache scales (#22818)
X-Git-Tag: upstream/0.0.10438~1369
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=1d72d87349f1be2eb5a60a15a75673481af907d6;p=pkg%2Fggml%2Fsources%2Fllama.cpp

convert : fix RuntimeError when stripping FP8 KV-cache scales (#22818)

* convert : fix RuntimeError when stripping FP8 KV-cache scales

In ModelBase._generate_nvfp4_tensors the final cleanup loop iterates
self.model_tensors.keys() and calls del on the same dict, which raises
RuntimeError: dictionary changed size during iteration when a ModelOpt
NVFP4 model also has FP8 KV-cache scales (e.g. mmangkad/Qwen3.6-35B-A3B-NVFP4
and any modelopt config with kv_cache_quant_algo: FP8).

Wrap the keys view in list() so the deletions happen on a snapshot.

* re-add another accidentally removed list

---------

Co-authored-by: Sigbjørn Skjæret <redacted>
---

diff --git a/convert_hf_to_gguf.py b/convert_hf_to_gguf.py
index fb9bc2361..400bf2bc4 100755
--- a/convert_hf_to_gguf.py
+++ b/convert_hf_to_gguf.py
@@ -710,7 +710,7 @@ class ModelBase:
                 self._repack_nvfp4(name, weight, scale, scale2, input_scale)
 
         # Flush any remaining experts (fallback if n_experts was unknown)
-        for bid, proj_type in expert_blocks.keys():
+        for bid, proj_type in list(expert_blocks.keys()):
             self._flush_nvfp4_experts((bid, proj_type), expert_blocks, expert_scales, expert_input_scales, expert_shapes, bid, proj_type)
 
         # Remove consumed tensors so get_tensors/modify_tensors won't see them
@@ -718,7 +718,7 @@ class ModelBase:
             self.model_tensors.pop(name, None)
 
         # Remove any remaining unused auxiliary tensors
-        for name in self.model_tensors.keys():
+        for name in list(self.model_tensors.keys()):
             if name.endswith((".k_scale", ".v_scale")):
                 del self.model_tensors[name]