openSUSE
/

CVE-Backport-Qwen2.5-Coder-32B

Text Generation

patch-backporting

code-generation

Model card Files Files and versions

CVE-Backport-Qwen2.5-Coder-32B

111 GB

Ctrl+K

Ctrl+K

1 contributor

History: 15 commits

anicka's picture

Add v5 LoRA adapter (QLoRA r=64 alpha=128, 36k codegen-only, 93.1% recall)

50eb0d3 verified 8 days ago

eval
Add eval results: eval/test-generation-v4-n50.json 14 days ago
v3-lora-adapter
Add v3 QLoRA adapter (r=64, 2.1GB, recommended for production) 13 days ago
v4-lora-adapter
Add v4 QLoRA adapter (r=64, 2.1GB) for further fine-tuning 14 days ago
v5-lora-adapter
Add v5 LoRA adapter (QLoRA r=64 alpha=128, 36k codegen-only, 93.1% recall) 8 days ago
.gitattributes

1.93 kB
Add v5 LoRA adapter (QLoRA r=64 alpha=128, 36k codegen-only, 93.1% recall) 8 days ago
Modelfile

596 Bytes
Add ollama Modelfile 15 days ago
README.md

4.94 kB
Update to v5: 93.1% recall, codegen-only dataset, adapted recall 90% 8 days ago
cve-backport-codegen-v3-q8_0.gguf

34.8 GB
xet

Add v3 Q8_0 GGUF (98% precision, no prompt echo — recommended for production) 13 days ago
cve-backport-codegen-v4-q8_0.gguf

34.8 GB
xet

Add v4 Q8_0 GGUF (33GB) 15 days ago
cve-backport-codegen-v5-q8_0.gguf

34.8 GB
xet

Add v5 Q8_0 GGUF (93.1% recall, codegen-only) 8 days ago