marlova/gemma-2-2B-it-thinking-function_calling-V0

README.md

@@ -34,7 +34,7 @@ This model was trained with SFT.
 
 ### Framework versions
 
-- TRL: 0.15.1
+- TRL: 0.15.2
 - Transformers: 4.48.3
 - Pytorch: 2.5.1+cu124
 - Datasets: 3.3.2

adapter_config.json

@@ -23,15 +23,15 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "k_proj",
+    "lm_head",
     "up_proj",
     "embed_tokens",
+    "o_proj",
+    "down_proj",
+    "k_proj",
     "v_proj",
     "q_proj",
-    "lm_head",
-    "gate_proj",
-    "down_proj"
+    "gate_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

runs/Feb26_13-35-29_36f309436a6c/events.out.tfevents.1740577045.36f309436a6c.4170.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a28e076ca88f104d1ef4b00ffe140280f7dd3d06b31b8d41326f9c30e63f61f1
+size 35396

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b7c95aadc8f74ec52eaa5c4b0b051754e9ec667397d2af61039e9a3f3119dae
+oid sha256:e939e1d15f20ad405315eb85ab4b33e9a44bf90830bd63ba862595e7d40ea94b
 size 5624