epimem: One-shot gradient-free learning on frozen transformers

- paper.md: full paper (Tommi Niemi / Rotko Networks)
- python/epimem.py: standalone Python reproduction
- export_onnx.py: ONNX export from HuggingFace (generates model files)
- results/memory_bank.json: example hidden-state vectors (896-dim)
- schema/: FlatBuffer schemas for memory bank + organism
- models/tokenizer/: Qwen 2.5 tokenizer files

Run: pip install transformers torch && python python/epimem.py
(Downloads Qwen 2.5 automatically from HuggingFace)

models/tokenizer/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}