- added (missing) Gemma3 bullet point in parent folder's readme.md (#788)

- typo in nbs

Author: casinca

ch05/12_gemma3/standalone-gemma3-plus-kvcache.ipynb

@@ -427,7 +427,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": null,
    "id": "e88de3e3-9f07-42cc-816b-28dbd46e96c4",
    "metadata": {
     "id": "e88de3e3-9f07-42cc-816b-28dbd46e96c4"
@@ -451,7 +451,7 @@
     "        self.cfg = cfg\n",
     "        self.current_pos = 0  # Track current position in KV cache\n",
     "\n",
-    "        # Reusuable utilities\n",
+    "        # Reusable utilities\n",
     "        cos_local, sin_local = compute_rope_params(\n",
     "            head_dim=cfg[\"head_dim\"],\n",
     "            theta_base=cfg[\"rope_local_base\"],\n",

ch05/12_gemma3/standalone-gemma3.ipynb

@@ -386,7 +386,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
+   "execution_count": null,
    "id": "e88de3e3-9f07-42cc-816b-28dbd46e96c4",
    "metadata": {
     "id": "e88de3e3-9f07-42cc-816b-28dbd46e96c4"
@@ -409,7 +409,7 @@
     "        self.out_head = nn.Linear(cfg[\"emb_dim\"], cfg[\"vocab_size\"], bias=False, dtype=cfg[\"dtype\"])\n",
     "        self.cfg = cfg\n",
     "\n",
-    "        # Reusuable utilities    \n",
+    "        # Reusable utilities    \n",
     "        cos_local, sin_local = compute_rope_params(\n",
     "            head_dim=cfg[\"head_dim\"],\n",
     "            theta_base=cfg[\"rope_local_base\"],\n",
@@ -1176,7 +1176,7 @@
    "provenance": []
   },
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": ".venv",
    "language": "python",
    "name": "python3"
   },
@@ -1190,7 +1190,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.16"
+   "version": "3.12.6"
   }
  },
  "nbformat": 4,

ch05/README.md

@@ -18,6 +18,7 @@
 - [09_extending-tokenizers](09_extending-tokenizers) contains a from-scratch implementation of the GPT-2 BPE tokenizer
 - [10_llm-training-speed](10_llm-training-speed) shows PyTorch performance tips to improve the LLM training speed
 - [11_qwen3](11_qwen3) A from-scratch implementation of Qwen3 0.6B and Qwen3 30B-A3B (Mixture-of-Experts) including code to load the pretrained weights of the base, reasoning, and coding model variants
+- [12_gemma3](12_gemma3) A from-scratch implementation of Gemma 3 270M and alternative with KV cache, including code to load the pretrained weights