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

README.md

@@ -36,7 +36,7 @@ This model was trained with SFT.
 
 - TRL: 0.17.0
 - Transformers: 4.51.3
-- Pytorch: 2.6.0+cu124
+- Pytorch: 2.7.0
 - Datasets: 3.5.1
 - Tokenizers: 0.21.1
 

adapter_config.json

@@ -24,15 +24,15 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "up_proj",
     "gate_proj",
-    "lm_head",
-    "k_proj",
-    "embed_tokens",
     "q_proj",
-    "up_proj",
-    "down_proj",
+    "k_proj",
+    "v_proj",
+    "lm_head",
     "o_proj",
-    "v_proj"
+    "embed_tokens",
+    "down_proj"
   ],
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3835634adf451edf5e87c39a150bfe53dc853d93f4e8f5862234a04722cd7543
+oid sha256:5e91f437056d3c48563edc6dbda1a10ca3579115561497a9b61376e36c249c88
 size 2475549872

runs/May07_14-44-48_MacBook-Pro.local/events.out.tfevents.1746654565.MacBook-Pro.local

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6456589cf94ad06093652088422124b5df348299d52966bfd466f10c0d33644c
+size 41027

special_tokens_map.json

@@ -18,7 +18,13 @@
     "rstrip": false,
     "single_word": false
   },
-  "eos_token": "<eos>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
   "pad_token": {
     "content": "<pad>",
     "lstrip": false,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b7395a8eba0b77ea630006f44d7b11b8cca7d9d2c2d6b3837e6a7462fe87b774
-size 5752
+oid sha256:58629d6973ce5eec2684405601b5742fdaa07a700f7894637951122d9496a65a
+size 6161