Model commit

config.json

@@ -0,0 +1,25 @@
+{
+  "model_type": "cosmicfish",
+  "architectures": [
+    "CosmicFish"
+  ],
+  "vocab_size": 50257,
+  "n_embd": 640,
+  "n_layer": 10,
+  "n_head": 16,
+  "block_size": 512,
+  "bias": true,
+  "dropout": 0.1,
+  "eps": 1e-06,
+  "use_rotary": true,
+  "use_swiglu": true,
+  "use_gqa": true,
+  "use_qk_norm": false,
+  "n_query_groups": 4,
+  "torch_dtype": "float16",
+  "transformers_version": "4.36.0",
+  "use_cache": true,
+  "pad_token_id": 50256,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c903c6f090dba21e3405a34c1005365d2a91b28ae72c494ec660b73f31b9e726
+size 183325771

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "tokenizer_class": "GPT2Tokenizer",
+  "vocab_size": 50257,
+  "model_max_length": 512,
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>",
+  "pad_token": "<|endoftext|>",
+  "add_prefix_space": false,
+  "do_lower_case": false
+}

vocab_info.json

@@ -0,0 +1,5 @@
+{
+  "note": "This model uses GPT-2 tokenizer. Please use: tokenizer = GPT2Tokenizer.from_pretrained('gpt2')",
+  "vocab_size": 50257,
+  "encoding": "gpt2"
+}