Training in progress, step 1000

config.json

@@ -1,5 +1,4 @@
 {
-  "_name_or_path": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
   "architectures": [
     "BitLlamaForCausalLM"
   ],
@@ -9,21 +8,21 @@
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
-  "hidden_size": 2048,
+  "hidden_size": 768,
   "initializer_range": 0.02,
-  "intermediate_size": 5632,
+  "intermediate_size": 2048,
   "max_position_embeddings": 2048,
   "model_type": "bit_llama",
-  "num_attention_heads": 32,
-  "num_hidden_layers": 22,
-  "num_key_value_heads": 4,
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "num_key_value_heads": 1,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "bfloat16",
+  "torch_dtype": "float32",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": true,
+  "use_cache": false,
   "vocab_size": 32002
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8c54963b4ac54779cd0d017c620f11d9be54fd744b724b68336883377e1c3049
-size 2201749928
+oid sha256:9af3f72ef755edb9973789511fd95f974b82de767a5b3c2420c6b4218a43a888
+size 485201912

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:07cc40c5a6d24066e3c9c9dc23db4ba29b9036550b49d4e3f00420e402d9f406
+oid sha256:2bf2c00a059d3cb812a18f7da5d445347629fae46c36e3a77776a100d50d2850
 size 4984