nvidia
/

Minitron-8B-Base

Text Generation

Model card Files Files and versions

Replace kv_channels with head_dim

#1

by mgoin - opened Jul 23, 2024

base: refs/heads/main

←

from: refs/pr/1

Discussion Files changed

Files changed (1) hide show

config.json +1 -1

config.json CHANGED Viewed

@@ -13,7 +13,7 @@
   "model_type": "nemotron",
   "num_attention_heads": 48,
   "num_hidden_layers": 32,
-  "kv_channels": 128,
   "num_key_value_heads": 8,
   "norm_eps": 1e-05,
   "rope_theta": 10000,

   "model_type": "nemotron",
   "num_attention_heads": 48,
   "num_hidden_layers": 32,
+  "head_dim": 128,
   "num_key_value_heads": 8,
   "norm_eps": 1e-05,
   "rope_theta": 10000,