Training in progress, epoch 1

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: meta-llama/Llama-3.1-8B
+base_model: meta-llama/Llama-3.2-3B
 library_name: transformers
 model_name: trained_model
 tags:
@@ -11,7 +11,7 @@ licence: license
 
 # Model Card for trained_model
 
-This model is a fine-tuned version of [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B).
+This model is a fine-tuned version of [meta-llama/Llama-3.2-3B](https://huggingface.co/meta-llama/Llama-3.2-3B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -36,7 +36,7 @@ This model was trained with SFT.
 
 - TRL: 0.19.1
 - Transformers: 4.53.2
-- Pytorch: 2.7.1+cu126
+- Pytorch: 2.7.1
 - Datasets: 4.0.0
 - Tokenizers: 0.21.2
 

config.json

@@ -8,27 +8,27 @@
   "eos_token_id": 128257,
   "head_dim": 128,
   "hidden_act": "silu",
-  "hidden_size": 4096,
+  "hidden_size": 3072,
   "initializer_range": 0.02,
-  "intermediate_size": 14336,
+  "intermediate_size": 8192,
   "max_position_embeddings": 131072,
   "mlp_bias": false,
   "model_type": "llama",
-  "num_attention_heads": 32,
-  "num_hidden_layers": 32,
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
   "num_key_value_heads": 8,
   "pad_token_id": 128257,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,
   "rope_scaling": {
-    "factor": 8.0,
+    "factor": 32.0,
     "high_freq_factor": 4.0,
     "low_freq_factor": 1.0,
     "original_max_position_embeddings": 8192,
     "rope_type": "llama3"
   },
   "rope_theta": 500000.0,
-  "tie_word_embeddings": false,
+  "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.53.2",
   "use_cache": true,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:519fbb2e7688c6aaa59ccb5881f4f70cbdea8f52064ec17e0cfb96eb8662462e
-size 8721
+oid sha256:f7b130f9bd80bc13e2a9d194d134b73a6dc9d54ba20909cf646809a7bb0e4a60
+size 7441