highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-09-14 18:51:53 +02:00
Code Issues Packages Projects Releases Wiki Activity

fixes for maverick

Browse Source
This commit is contained in:
Michael Yang
2025-04-21 10:45:56 -07:00
committed by Michael Yang
parent 8bf11b84c1
commit 7ba9fa9c7d
3 changed files with 53 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
convert/convert_llama4.go
View File

@@ -45,8 +45,8 @@ func (p *llama4Model) KV(t *Tokenizer) ggml.KV {
}
}
kv["llama4.intermediate_size"] = p.TextModel.IntermediateSizeMLP
kv["llama4.intermediate_size_moe"] = p.TextModel.IntermediateSize
kv["llama4.feed_forward_length"] = p.TextModel.IntermediateSizeMLP
kv["llama4.expert_feed_forward_length"] = p.TextModel.IntermediateSize
kv["llama4.expert_count"] = p.TextModel.NumLocalExperts
kv["llama4.expert_used_count"] = p.TextModel.NumExpertsPerToken