Upload folder using huggingface_hub

README.md

@@ -8,11 +8,11 @@ Chain of Thought (CoT) transformer model trained to do multi-step integer arithm
 
 Model details:
  - **Vocabulary Size**: 40 (Character Tokenization)
- - **Layer Count**: 8
+ - **Layer Count**: 12
  - **Attention Head Count**: 4
- - **Residual Stream Size**: 256
+ - **Residual Stream Size**: 512
  - **Context Length**: 256
- - **Tokens Trained on**: 419,612,160
+ - **Tokens Trained on**: 1,214,656,000
 
 Training Score During Training
 

config.json

@@ -4,17 +4,18 @@
   ],
   "attention_dropout": 0.0,
   "bos_token_id": null,
+  "dtype": "float32",
   "embd_pdrop": 0.0,
   "eos_token_id": null,
   "hidden_act": "gelu_new",
-  "hidden_size": 256,
+  "hidden_size": 512,
   "initializer_range": 0.02,
-  "intermediate_size": 1024,
+  "intermediate_size": 2048,
   "layer_norm_eps": 1e-05,
   "max_position_embeddings": 256,
   "model_type": "phi",
   "num_attention_heads": 4,
-  "num_hidden_layers": 8,
+  "num_hidden_layers": 12,
   "num_key_value_heads": 4,
   "partial_rotary_factor": 0.5,
   "qk_layernorm": false,
@@ -22,8 +23,7 @@
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "float32",
-  "transformers_version": "4.53.0",
+  "transformers_version": "4.57.1",
   "use_cache": true,
-  "vocab_size": 40
+  "vocab_size": 45
 }

generation_config.json

@@ -1,4 +1,4 @@
 {
   "_from_model_config": true,
-  "transformers_version": "4.53.0"
+  "transformers_version": "4.57.1"
 }

hyperparameters.json

@@ -1 +1 @@
-{"MIN_DIFFICULTY": 2, "MAX_DIFFICULTY": 4, "TRAINING_SAMPLES": 3000000, "CONTEXT_LENGTH": 256, "RESIDUAL_EMBEDDING_SIZE": 256, "MLP_EMBEDDING_SIZE": 1024, "NUM_ATTENTION_HEADS": 4, "NUM_LAYERS": 8, "VOCAB_SIZE": 40, "TOTAL_TOKENS": 419612160}
+{"MIN_DIFFICULTY": 2, "MAX_DIFFICULTY": 6, "TRAINING_SAMPLES": 3000000, "CONTEXT_LENGTH": 256, "RESIDUAL_EMBEDDING_SIZE": 512, "MLP_EMBEDDING_SIZE": 2048, "NUM_ATTENTION_HEADS": 4, "NUM_LAYERS": 12, "VOCAB_SIZE": 40, "TOTAL_TOKENS": 1214656000, "EPOCHS": 2}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:80470ec0e1be3e0e2ae474707fdd150d02d666915dfaa5d54528418c03e7da66
-size 25352072
+oid sha256:84cbed7983068ea560ef90a7f02e1408c244eb4e18315d3640ff1cd0114f6f80
+size 151471932

tokenizer.json

@@ -115,9 +115,9 @@
       "Ġ": 39,
       "Ġ-": 40,
       "(-": 41,
-      "Ġ1": 42,
-      "St": 43,
-      "ep": 44
+      "Ġ+": 42,
+      "Ġ*": 43,
+      "Ġ1": 44
     },
     "merges": [
       [
@@ -130,15 +130,15 @@
       ],
       [
         "Ġ",
-        "1"
+        "+"
       ],
       [
-        "S",
-        "t"
+        "Ġ",
+        "*"
       ],
       [
-        "e",
-        "p"
+        "Ġ",
+        "1"
       ]
     ]
   }