Training in progress, step 500

adapter_config.json

@@ -16,7 +16,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 32,
+  "lora_alpha": 128,
   "lora_bias": false,
   "lora_dropout": 0.0,
   "megatron_config": null,
@@ -25,17 +25,17 @@
   "peft_type": "LORA",
   "peft_version": "0.18.0",
   "qalora_group_size": 16,
-  "r": 16,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "q_proj",
-    "gate_proj",
     "k_proj",
-    "v_proj",
+    "gate_proj",
     "down_proj",
-    "up_proj"
+    "v_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4e3ef301cb916cd9842e7bcc7bedddc9f3312b77001984289ac593bf45fc0c7
-size 40422168
+oid sha256:933af8dc6e170a6b7ba5c74f7e681f66dd9649396b68f08e8f30626b17d82eeb
+size 161533160

config.json

@@ -161,9 +161,9 @@
   "label_smoothing": 0.0,
   "length_penalty": 1.0,
   "llm_dim": 1024,
-  "lora_alpha": 32,
+  "lora_alpha": 128,
   "lora_dropout": 0.0,
-  "lora_rank": 16,
+  "lora_rank": 64,
   "lora_target_modules": [
     "q_proj",
     "k_proj",

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5682a2fc840b85276585ed818f9c67c64a20a8dc3acc70e761134325176d27a
+oid sha256:770dc58ad85229de4f3ef07623e96976c58989ff8573bbc04d5302ba909f540d
 size 25172384