Training in progress, step 1000

config.json

@@ -4,20 +4,19 @@
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "bos_token_id": 200000,
-  "eos_token_id": 201136,
-  "head_dim": 128,
+  "bos_token_id": 128256,
+  "eos_token_id": 128257,
+  "head_dim": 64,
   "hidden_act": "silu",
-  "hidden_size": 1024,
+  "hidden_size": 576,
   "initializer_range": 0.02,
-  "intermediate_size": 2752,
+  "intermediate_size": 1536,
   "max_position_embeddings": 2048,
   "mlp_bias": false,
   "model_type": "llama",
-  "num_attention_heads": 8,
-  "num_hidden_layers": 12,
-  "num_key_value_heads": 8,
-  "pad_token_id": 200018,
+  "num_attention_heads": 9,
+  "num_hidden_layers": 30,
+  "num_key_value_heads": 3,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,
@@ -26,5 +25,5 @@
   "torch_dtype": "float32",
   "transformers_version": "4.51.3",
   "use_cache": false,
-  "vocab_size": 201145
+  "vocab_size": 128268
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e198627cafd4909f26df24b47da1232b6a3ade8aa35236ad2a5215236a13162
-size 2255020440
+oid sha256:fd0a8f14efc2f0a49f278cf4e8cff1666bcf2acf6cad8d1c11de34c006346398
+size 1015903256

special_tokens_map.json

@@ -2,6 +2,7 @@
   "additional_special_tokens": [
     "<|im_start|>",
     "<|im_end|>",
+    "<|finetune_pad|>",
     "<tools>",
     "</tools>",
     "<tool_call>",
@@ -12,7 +13,7 @@
     "</think>"
   ],
   "bos_token": {
-    "content": "<|begin_of_text|>",
+    "content": "<|im_start|>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,
@@ -25,11 +26,5 @@
     "rstrip": false,
     "single_word": false
   },
-  "pad_token": {
-    "content": "<|finetune_right_pad|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
+  "pad_token": "<finetune_pad>"
 }

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:be73bc5046bc7cc697ed444d8ebd4e4b8d0452b72fcc921521f075b303071d78
-size 27950454
+oid sha256:9e3ed6c2fb0cdfe4d06a4e33d02ba6c0a99b0afa85d0f8a8a50527aed2d1767d
+size 17212221

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e9de1c3be660bcc9e2ce5fe5d97a2e4bb4a46f3f275ef49426b9abd7097f4349
-size 6840
+oid sha256:dc85aa028e904531eb3094cca8ea7d5c1b0f768d5faeeb4854ca98a752ec3db5
+size 6904