Add gpt2 base mlx weights

README.md

@@ -0,0 +1,10 @@
+---
+language: en
+tags:
+- exbert
+- mlx
+license: mit
+pipeline_tag: text-generation
+base_model: openai-community/gpt2
+library_name: mlx
+---

config.json

@@ -0,0 +1,31 @@
+{
+    "activation_function": "gelu_new",
+    "architectures": [
+        "GPT2LMHeadModel"
+    ],
+    "attn_pdrop": 0.1,
+    "bos_token_id": 50256,
+    "embd_pdrop": 0.1,
+    "eos_token_id": 50256,
+    "initializer_range": 0.02,
+    "layer_norm_epsilon": 1e-05,
+    "model_type": "gpt2",
+    "n_ctx": 1024,
+    "n_embd": 768,
+    "n_head": 12,
+    "n_layer": 12,
+    "n_positions": 1024,
+    "resid_pdrop": 0.1,
+    "summary_activation": null,
+    "summary_first_dropout": 0.1,
+    "summary_proj_to_labels": true,
+    "summary_type": "cls_index",
+    "summary_use_proj": true,
+    "task_specific_params": {
+        "text-generation": {
+            "do_sample": true,
+            "max_length": 50
+        }
+    },
+    "vocab_size": 50257
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "transformers_version": "4.26.0.dev0",
+  "_from_model_config": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b48ea073b280cd35b7d19c4ee91656db2a203cd3a84f91313fe350ac96f7c8f0
+size 497773308

model.safetensors.index.json

@@ -0,0 +1,156 @@
+{
+    "metadata": {
+        "total_size": 497759232,
+        "total_parameters": 124439808
+    },
+    "weight_map": {
+        "model.h.0.attn.c_attn.bias": "model.safetensors",
+        "model.h.0.attn.c_attn.weight": "model.safetensors",
+        "model.h.0.attn.c_proj.bias": "model.safetensors",
+        "model.h.0.attn.c_proj.weight": "model.safetensors",
+        "model.h.0.ln_1.bias": "model.safetensors",
+        "model.h.0.ln_1.weight": "model.safetensors",
+        "model.h.0.ln_2.bias": "model.safetensors",
+        "model.h.0.ln_2.weight": "model.safetensors",
+        "model.h.0.mlp.c_fc.bias": "model.safetensors",
+        "model.h.0.mlp.c_fc.weight": "model.safetensors",
+        "model.h.0.mlp.c_proj.bias": "model.safetensors",
+        "model.h.0.mlp.c_proj.weight": "model.safetensors",
+        "model.h.1.attn.c_attn.bias": "model.safetensors",
+        "model.h.1.attn.c_attn.weight": "model.safetensors",
+        "model.h.1.attn.c_proj.bias": "model.safetensors",
+        "model.h.1.attn.c_proj.weight": "model.safetensors",
+        "model.h.1.ln_1.bias": "model.safetensors",
+        "model.h.1.ln_1.weight": "model.safetensors",
+        "model.h.1.ln_2.bias": "model.safetensors",
+        "model.h.1.ln_2.weight": "model.safetensors",
+        "model.h.1.mlp.c_fc.bias": "model.safetensors",
+        "model.h.1.mlp.c_fc.weight": "model.safetensors",
+        "model.h.1.mlp.c_proj.bias": "model.safetensors",
+        "model.h.1.mlp.c_proj.weight": "model.safetensors",
+        "model.h.10.attn.c_attn.bias": "model.safetensors",
+        "model.h.10.attn.c_attn.weight": "model.safetensors",
+        "model.h.10.attn.c_proj.bias": "model.safetensors",
+        "model.h.10.attn.c_proj.weight": "model.safetensors",
+        "model.h.10.ln_1.bias": "model.safetensors",
+        "model.h.10.ln_1.weight": "model.safetensors",
+        "model.h.10.ln_2.bias": "model.safetensors",
+        "model.h.10.ln_2.weight": "model.safetensors",
+        "model.h.10.mlp.c_fc.bias": "model.safetensors",
+        "model.h.10.mlp.c_fc.weight": "model.safetensors",
+        "model.h.10.mlp.c_proj.bias": "model.safetensors",
+        "model.h.10.mlp.c_proj.weight": "model.safetensors",
+        "model.h.11.attn.c_attn.bias": "model.safetensors",
+        "model.h.11.attn.c_attn.weight": "model.safetensors",
+        "model.h.11.attn.c_proj.bias": "model.safetensors",
+        "model.h.11.attn.c_proj.weight": "model.safetensors",
+        "model.h.11.ln_1.bias": "model.safetensors",
+        "model.h.11.ln_1.weight": "model.safetensors",
+        "model.h.11.ln_2.bias": "model.safetensors",
+        "model.h.11.ln_2.weight": "model.safetensors",
+        "model.h.11.mlp.c_fc.bias": "model.safetensors",
+        "model.h.11.mlp.c_fc.weight": "model.safetensors",
+        "model.h.11.mlp.c_proj.bias": "model.safetensors",
+        "model.h.11.mlp.c_proj.weight": "model.safetensors",
+        "model.h.2.attn.c_attn.bias": "model.safetensors",
+        "model.h.2.attn.c_attn.weight": "model.safetensors",
+        "model.h.2.attn.c_proj.bias": "model.safetensors",
+        "model.h.2.attn.c_proj.weight": "model.safetensors",
+        "model.h.2.ln_1.bias": "model.safetensors",
+        "model.h.2.ln_1.weight": "model.safetensors",
+        "model.h.2.ln_2.bias": "model.safetensors",
+        "model.h.2.ln_2.weight": "model.safetensors",
+        "model.h.2.mlp.c_fc.bias": "model.safetensors",
+        "model.h.2.mlp.c_fc.weight": "model.safetensors",
+        "model.h.2.mlp.c_proj.bias": "model.safetensors",
+        "model.h.2.mlp.c_proj.weight": "model.safetensors",
+        "model.h.3.attn.c_attn.bias": "model.safetensors",
+        "model.h.3.attn.c_attn.weight": "model.safetensors",
+        "model.h.3.attn.c_proj.bias": "model.safetensors",
+        "model.h.3.attn.c_proj.weight": "model.safetensors",
+        "model.h.3.ln_1.bias": "model.safetensors",
+        "model.h.3.ln_1.weight": "model.safetensors",
+        "model.h.3.ln_2.bias": "model.safetensors",
+        "model.h.3.ln_2.weight": "model.safetensors",
+        "model.h.3.mlp.c_fc.bias": "model.safetensors",
+        "model.h.3.mlp.c_fc.weight": "model.safetensors",
+        "model.h.3.mlp.c_proj.bias": "model.safetensors",
+        "model.h.3.mlp.c_proj.weight": "model.safetensors",
+        "model.h.4.attn.c_attn.bias": "model.safetensors",
+        "model.h.4.attn.c_attn.weight": "model.safetensors",
+        "model.h.4.attn.c_proj.bias": "model.safetensors",
+        "model.h.4.attn.c_proj.weight": "model.safetensors",
+        "model.h.4.ln_1.bias": "model.safetensors",
+        "model.h.4.ln_1.weight": "model.safetensors",
+        "model.h.4.ln_2.bias": "model.safetensors",
+        "model.h.4.ln_2.weight": "model.safetensors",
+        "model.h.4.mlp.c_fc.bias": "model.safetensors",
+        "model.h.4.mlp.c_fc.weight": "model.safetensors",
+        "model.h.4.mlp.c_proj.bias": "model.safetensors",
+        "model.h.4.mlp.c_proj.weight": "model.safetensors",
+        "model.h.5.attn.c_attn.bias": "model.safetensors",
+        "model.h.5.attn.c_attn.weight": "model.safetensors",
+        "model.h.5.attn.c_proj.bias": "model.safetensors",
+        "model.h.5.attn.c_proj.weight": "model.safetensors",
+        "model.h.5.ln_1.bias": "model.safetensors",
+        "model.h.5.ln_1.weight": "model.safetensors",
+        "model.h.5.ln_2.bias": "model.safetensors",
+        "model.h.5.ln_2.weight": "model.safetensors",
+        "model.h.5.mlp.c_fc.bias": "model.safetensors",
+        "model.h.5.mlp.c_fc.weight": "model.safetensors",
+        "model.h.5.mlp.c_proj.bias": "model.safetensors",
+        "model.h.5.mlp.c_proj.weight": "model.safetensors",
+        "model.h.6.attn.c_attn.bias": "model.safetensors",
+        "model.h.6.attn.c_attn.weight": "model.safetensors",
+        "model.h.6.attn.c_proj.bias": "model.safetensors",
+        "model.h.6.attn.c_proj.weight": "model.safetensors",
+        "model.h.6.ln_1.bias": "model.safetensors",
+        "model.h.6.ln_1.weight": "model.safetensors",
+        "model.h.6.ln_2.bias": "model.safetensors",
+        "model.h.6.ln_2.weight": "model.safetensors",
+        "model.h.6.mlp.c_fc.bias": "model.safetensors",
+        "model.h.6.mlp.c_fc.weight": "model.safetensors",
+        "model.h.6.mlp.c_proj.bias": "model.safetensors",
+        "model.h.6.mlp.c_proj.weight": "model.safetensors",
+        "model.h.7.attn.c_attn.bias": "model.safetensors",
+        "model.h.7.attn.c_attn.weight": "model.safetensors",
+        "model.h.7.attn.c_proj.bias": "model.safetensors",
+        "model.h.7.attn.c_proj.weight": "model.safetensors",
+        "model.h.7.ln_1.bias": "model.safetensors",
+        "model.h.7.ln_1.weight": "model.safetensors",
+        "model.h.7.ln_2.bias": "model.safetensors",
+        "model.h.7.ln_2.weight": "model.safetensors",
+        "model.h.7.mlp.c_fc.bias": "model.safetensors",
+        "model.h.7.mlp.c_fc.weight": "model.safetensors",
+        "model.h.7.mlp.c_proj.bias": "model.safetensors",
+        "model.h.7.mlp.c_proj.weight": "model.safetensors",
+        "model.h.8.attn.c_attn.bias": "model.safetensors",
+        "model.h.8.attn.c_attn.weight": "model.safetensors",
+        "model.h.8.attn.c_proj.bias": "model.safetensors",
+        "model.h.8.attn.c_proj.weight": "model.safetensors",
+        "model.h.8.ln_1.bias": "model.safetensors",
+        "model.h.8.ln_1.weight": "model.safetensors",
+        "model.h.8.ln_2.bias": "model.safetensors",
+        "model.h.8.ln_2.weight": "model.safetensors",
+        "model.h.8.mlp.c_fc.bias": "model.safetensors",
+        "model.h.8.mlp.c_fc.weight": "model.safetensors",
+        "model.h.8.mlp.c_proj.bias": "model.safetensors",
+        "model.h.8.mlp.c_proj.weight": "model.safetensors",
+        "model.h.9.attn.c_attn.bias": "model.safetensors",
+        "model.h.9.attn.c_attn.weight": "model.safetensors",
+        "model.h.9.attn.c_proj.bias": "model.safetensors",
+        "model.h.9.attn.c_proj.weight": "model.safetensors",
+        "model.h.9.ln_1.bias": "model.safetensors",
+        "model.h.9.ln_1.weight": "model.safetensors",
+        "model.h.9.ln_2.bias": "model.safetensors",
+        "model.h.9.ln_2.weight": "model.safetensors",
+        "model.h.9.mlp.c_fc.bias": "model.safetensors",
+        "model.h.9.mlp.c_fc.weight": "model.safetensors",
+        "model.h.9.mlp.c_proj.bias": "model.safetensors",
+        "model.h.9.mlp.c_proj.weight": "model.safetensors",
+        "model.ln_f.bias": "model.safetensors",
+        "model.ln_f.weight": "model.safetensors",
+        "model.wpe.weight": "model.safetensors",
+        "model.wte.weight": "model.safetensors"
+    }
+}

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,20 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "50256": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1024,
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}