Upload tiny random Llama-3.3 model (seed 42)

README.md

@@ -24,16 +24,16 @@ This is a tiny random version of the meta-llama/Llama-3.3-70B-Instruct model, cr
 
 ## Parameters
 
-- **Total parameters**: ~75,110,656
-- **Trainable parameters**: ~75,110,656
+- **Total parameters**: ~42,277,376
+- **Trainable parameters**: ~42,277,376
 
 ## Usage
 
 ```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import AutoModelForSequenceClassification, AutoTokenizer
 
 # Load model and tokenizer
-model = AutoModelForCausalLM.from_pretrained("AlignmentResearch/Llama-3.3-Tiny-Instruct")
+model = AutoModelForSequenceClassification.from_pretrained("AlignmentResearch/Llama-3.3-Tiny-Instruct")
 tokenizer = AutoTokenizer.from_pretrained("AlignmentResearch/Llama-3.3-Tiny-Instruct")
 
 # Generate text (note: this model has random weights!)

config.json

@@ -1,6 +1,6 @@
 {
   "architectures": [
-    "LlamaForCausalLM"
+    "LlamaForSequenceClassification"
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
@@ -13,8 +13,14 @@
   "head_dim": 64,
   "hidden_act": "silu",
   "hidden_size": 256,
+  "id2label": {
+    "0": "LABEL_0"
+  },
   "initializer_range": 0.02,
   "intermediate_size": 768,
+  "label2id": {
+    "LABEL_0": 0
+  },
   "max_position_embeddings": 131072,
   "mlp_bias": false,
   "model_type": "llama",
@@ -32,7 +38,7 @@
   },
   "rope_theta": 500000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "float32",
+  "torch_dtype": "bfloat16",
   "transformers_version": "4.54.1",
   "use_cache": true,
   "vocab_size": 128256

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90be80471e3b86e6f1aaea7c053977f9aa0f720669ca9635b5f97bba661a0f2c
-size 300454912
+oid sha256:aadd7211b5c104ff62695304c0dbf8884790d7e99a148d2d2ca57b7fd077bdd4
+size 84567424