Training in progress, step 25

adapter_config.json

@@ -7,21 +7,20 @@
   "init_lora_weights": true,
   "layers_pattern": null,
   "layers_to_transform": null,
-  "lora_alpha": 32,
+  "lora_alpha": 16,
   "lora_dropout": 0.05,
   "modules_to_save": [
     "connector",
     "mm_projector",
     "response_head",
-    "related_head",
     "lm_head",
     "informative_head",
     "relevance_head",
     "uncertainty_head"
   ],
   "peft_type": "LORA",
-  "r": 16,
+  "r": 8,
   "revision": null,
-  "target_modules": "model\\.layers.*(q_proj|k_proj|v_proj|o_proj|gate_proj|up_proj|down_proj)$",
+  "target_modules": "model\\.layers.*(q_proj|k_proj|v_proj|gate_proj)$",
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4b105f8ba9adeb788677a2b0fc9f3480c4afb4a2fc1b943d0cce350b32bf647
-size 1204780872
+oid sha256:c1ade9748556e4d56886cccd85bf923f95e6bfecdeaab4d697308aa9a23a157c
+size 1140991056

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9a8273c75b0094b53824030ce4878138db8805782a9abaa305b557281cfaf2bd
+oid sha256:5fe8b61c87f3a88ae769d377e2273c4fdb025ccd9b26bc8dd91440af90212fd2
 size 7672