Upload LoRA adapter (README written by author)

README.md

@@ -35,7 +35,7 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Method: QLoRA (4-bit)
 - Max sequence length: 2048
 - Epochs: 1
-- Learning rate: 1e-06
+- Learning rate: 2e-06
 - LoRA: r=64, alpha=128
 
 ## Usage

adapter_config.json

@@ -22,7 +22,7 @@
   "loftq_config": {},
   "lora_alpha": 128,
   "lora_bias": false,
-  "lora_dropout": 0.0,
+  "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
@@ -33,12 +33,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
     "v_proj",
     "o_proj",
-    "up_proj",
-    "q_proj",
     "gate_proj",
+    "q_proj",
+    "down_proj",
+    "up_proj",
     "k_proj"
   ],
   "target_parameters": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a9ef53ca843e4689c909d617e8ae37e6498f5323337aa49f1ba874b70a0800b
+oid sha256:369fc060e600597514aab2c540887fc53e04e96693120e2c89d2e98a8a079ed1
 size 528550256