Update model card with metadata tags

README.md

@@ -1,14 +1,28 @@
 ---
 license: apache-2.0
-library_name: pytorch
+library_name: peft
+pipeline_tag: text-generation
+base_model: tiiuae/Falcon3-10B-Base
 tags:
-  - checkpoints
   - qlora
+  - peft
   - falcon3
+  - checkpoints
   - post-training
-base_model: tiiuae/Falcon3-10B-Base
+  - scientific-assistant
+language:
+  - en
 ---
 
-# Nexa Checkpoints
+# Falcon3 QLoRA Checkpoints (Nexa)
+
+This repository contains exported checkpoint artifacts from `nexa_inference/checkpoints`.
+
+## Contents
+- QLoRA checkpoint directory: `nexa_inference/checkpoints/falcon3_qlora_v1/checkpoint-200/`
+- Tokenizer/config metadata needed to reload the adapter checkpoint.
 
-Checkpoint artifacts exported from `nexa_inference/checkpoints`.
+## Notes
+- Base model: `tiiuae/Falcon3-10B-Base`
+- Training method: QLoRA + PEFT adapters
+- Scope: checkpoint storage and reproducibility metadata