llamamodel: add DeepSeek-V2 to whitelist (nomic-ai#2702)

Signed-off-by: Jared Van Bortel <[email protected]>
mojowebs · Jul 22, 2024 · 4ca1d04 · 4ca1d04
1 parent 54ed309
commit 4ca1d04
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/gpt4all-backend/llamamodel.cpp b/gpt4all-backend/llamamodel.cpp
@@ -81,7 +81,7 @@ static const std::vector<const char *> KNOWN_ARCHES {
     "olmo",
     "openelm",
     // "arctic", -- 10B+128x3.66B parameters
-    // "deepseek2", -- excessive VRAM requirements
+    "deepseek2",
     "chatglm",
     // "bitnet", -- tensor not within file bounds?
     // "t5", -- seq2seq model