Training in progress, step 5

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "down_proj",
+    "v_proj",
     "up_proj",
+    "down_proj",
+    "o_proj",
     "k_proj",
     "q_proj",
-    "gate_proj",
-    "v_proj"
+    "gate_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 sha256:95acb38d014584c6b40597db7ee65c8089400fb21322fdadfc4f31ae2a16e3b5
-size 456206712
+oid sha256:39d4dd96080c3616aa6d08dae3d76f1b345376cc1812a392813a337b646bfcf0
+size 3140603848

config.json

@@ -4,7 +4,7 @@
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 1,
-  "eos_token_id": 15,
+  "eos_token_id": 131072,
   "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 5120,
@@ -37,5 +37,5 @@
   "torch_dtype": "bfloat16",
   "transformers_version": "4.52.4",
   "use_cache": false,
-  "vocab_size": 131072
+  "vocab_size": 131074
 }

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4b4c8fcd33487a449c07f423d47adb035bba8347ccf13eb074b4d1fef8acf919
-size 17078288
+oid sha256:a2fa2956478eaa353c6c4b1f47fdd6868cce6075e52e169c35ae8bd28524e7a8
+size 17078668

tokenizer_config.json

@@ -116,7 +116,7 @@
       "special": true
     },
     "14": {
-      "content": "<|im_start|>",
+      "content": "<SPECIAL_14>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -124,7 +124,7 @@
       "special": true
     },
     "15": {
-      "content": "<|im_end|>",
+      "content": "<SPECIAL_15>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -8002,6 +8002,22 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "131072": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "131073": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
     }
   },
   "bos_token": "<s>",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:00680aba3e8386e5d6534a13bd4e116c275cde571d5790cdf1205c8fcd88cefb
+oid sha256:b863d975b84ce444ea5c542d9fabff254c1ce04e88ce402c0259419976063185
 size 9489