Training in progress, step 4325

README.md

@@ -1,18 +1,18 @@
 ---
-base_model: Ba2han/kumru-model-merged
+base_model: vngrs-ai/Kumru-2B-Base
 library_name: transformers
 model_name: k-outputs-2
 tags:
 - generated_from_trainer
+- trl
 - sft
 - unsloth
-- trl
 licence: license
 ---
 
 # Model Card for k-outputs-2
 
-This model is a fine-tuned version of [Ba2han/kumru-model-merged](https://huggingface.co/Ba2han/kumru-model-merged).
+This model is a fine-tuned version of [vngrs-ai/Kumru-2B-Base](https://huggingface.co/vngrs-ai/Kumru-2B-Base).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -28,18 +28,18 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/batuhan409/huggingface/runs/e7zcc37u) 
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/batuhan409/huggingface/runs/f9s9q25z) 
 
 
 This model was trained with SFT.
 
 ### Framework versions
 
-- TRL: 0.21.0
-- Transformers: 4.56.1
-- Pytorch: 2.8.0
+- TRL: 0.22.2
+- Transformers: 4.55.4
+- Pytorch: 2.8.0+cu128
 - Datasets: 4.2.0
-- Tokenizers: 0.22.0
+- Tokenizers: 0.21.4
 
 ## Citations
 

config.json

@@ -4,8 +4,7 @@
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 2,
-  "dtype": "bfloat16",
-  "eos_token_id": 3,
+  "eos_token_id": 50177,
   "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 3072,
@@ -22,8 +21,9 @@
   "rope_theta": 500000,
   "sliding_window": null,
   "tie_word_embeddings": false,
-  "transformers_version": "4.56.1",
-  "unsloth_version": "2025.10.6",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.55.4",
+  "unsloth_version": "2025.10.8",
   "use_cache": true,
-  "vocab_size": 50176
+  "vocab_size": 50178
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec5f5a1d971331ec93c21e1a96fa4ee14b0caab3609c9140e89d4e67fd72cdcf
-size 4750295696
+oid sha256:dd293cb5a4e52ca2610ea73e8c25d357b8364ac2c79187f4048b5675c4899d07
+size 4750320272

special_tokens_map.json

@@ -1,4 +1,13 @@
 {
+  "additional_special_tokens": [
+    {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
   "bos_token": {
     "content": "<BOS>",
     "lstrip": false,
@@ -7,7 +16,7 @@
     "single_word": false
   },
   "eos_token": {
-    "content": "<EOS>",
+    "content": "<|im_end|>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer.json

@@ -2360,6 +2360,24 @@
       "rstrip": false,
       "normalized": false,
       "special": true
+    },
+    {
+      "id": 50176,
+      "content": "<|im_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 50177,
+      "content": "<|im_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
     }
   ],
   "normalizer": null,

tokenizer_config.json

@@ -2095,13 +2095,32 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "50176": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50177": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
     }
   },
+  "additional_special_tokens": [
+    "<|im_start|>"
+  ],
   "bos_token": "<BOS>",
   "clean_up_tokenization_spaces": true,
-  "eos_token": "<EOS>",
+  "eos_token": "<|im_end|>",
   "extra_special_tokens": {},
-  "model_max_length": 8192,
+  "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<PAD>",
   "padding_side": "right",
   "tokenizer_class": "PreTrainedTokenizerFast",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1a892b6dad659c28027919dc2489199621b8670210f52e1ce216af7b24775181
+oid sha256:96b29caca31af828c363f434339163116665b126cea9906ed02c22c78eb31622
 size 6225