docs: add canonical tooling corpus (147 files) from Google/HF/frameworks

Five-lane parallel research pass. Each subdir under tooling/ has its own
README indexing downloaded files with verified upstream sources.

- google-official/: deepmind-gemma JAX examples, gemma_pytorch scripts,
  gemma.cpp API server docs, google-gemma/cookbook notebooks, ai.google.dev
  HTML snapshots, Gemma 3 tech report
- huggingface/: 8 gemma-4-* model cards, chat-template .jinja files,
  tokenizer_config.json, transformers gemma4/ source, launch blog posts,
  official HF Spaces app.py
- inference-frameworks/: vLLM/llama.cpp/MLX/Keras-hub/TGI/Gemini API/Vertex AI
  comparison, run_commands.sh with 8 working launches, 9 code snippets
- gemma-family/: 12 per-variant briefs (ShieldGemma 2, CodeGemma, PaliGemma 2,
  Recurrent/Data/Med/TxGemma, Embedding/Translate/Function/Dolphin/SignGemma)
- fine-tuning/: Unsloth Gemma 4 notebooks, Axolotl YAMLs (incl 26B-A4B MoE),
  TRL scripts, Google cookbook fine-tune notebooks, recipe-recommendation.md

Findings that update earlier CORPUS_* docs are flagged in tooling/README.md
(not applied) — notably the new <|turn>/<turn|> prompt format, gemma_pytorch
abandonment, gemma.cpp Gemini-API server, transformers AutoModelForMultimodalLM,
FA2 head_dim=512 break, 26B-A4B MoE quantization rules, no Gemma 4 tech
report PDF yet, no Gemma-4-generation specialized siblings yet.

Pre-commit secrets hook bypassed per user authorization — flagged "secrets"
are base64 notebook cell outputs and example Ed25519 keys in the HDP
agentic-security demo, not real credentials.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

tooling/fine-tuning/axolotl/e2b-vision-lora.yaml

@@ -0,0 +1,62 @@
+# Gemma 4 E2B Vision LoRA
+#
+# Fine-tuning LM LoRA adapters on multimodal Gemma4 with vision/multimodal modules frozen.
+# Uses the base ProcessingStrategy (auto-detects image_token from processor).
+
+base_model: google/gemma-4-E2B-it
+processor_type: AutoProcessor
+freeze_mm_modules: true
+
+plugins:
+  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
+strict: false
+
+# Required for vision/multimodal training
+skip_prepare_dataset: true
+remove_unused_columns: false
+sample_packing: false
+
+chat_template: gemma4
+datasets:
+  - path: HuggingFaceH4/llava-instruct-mix-vsft
+    type: chat_template
+    split: train[:100]
+
+val_set_size: 0
+output_dir: ./outputs/gemma4-e2b-vision-lora
+
+adapter: lora
+sequence_len: 2048
+pad_to_sequence_len: false
+
+lora_r: 16
+lora_alpha: 32
+lora_dropout: 0
+# Target language model only — vision encoder is frozen via freeze_mm_modules
+lora_target_modules: 'model.language_model.layers.[\d]+.(_checkpoint_wrapped_module.)?(mlp|self_attn).(up|down|gate|q|k|v|o)_proj'
+
+gradient_accumulation_steps: 4
+micro_batch_size: 1
+num_epochs: 1
+max_steps: 10
+optimizer: adamw_torch_8bit
+lr_scheduler: cosine
+learning_rate: 0.0002
+
+bf16: auto
+tf32: true
+
+gradient_checkpointing: true
+gradient_checkpointing_kwargs:
+  use_reentrant: false
+logging_steps: 1
+sdp_attention: true
+
+warmup_ratio: 0.1
+weight_decay: 0.0
+
+wandb_project:
+wandb_entity:
+wandb_watch:
+wandb_name:
+wandb_log_model: