Update Ollama model configurations and documentation for clarity and consistency

- Revised model descriptions in `ollama-models.example.yaml` to emphasize uncensored and abliterated requirements, enhancing user understanding.
- Updated documentation in `llm.md` to reflect changes in model tags and their meanings, ensuring accurate guidance for users.
- Modified CLI help messages to clarify the nature of presets, reinforcing that all options are abliterate models with Russian support.
- Enhanced the `llm_runtime.py` file to align preset labels with the updated model descriptions, improving consistency across the codebase.

src/gpu_rent/llm_runtime.py

@@ -16,22 +16,20 @@ from gpu_rent.paths import (
 VALID_RUNTIMES = frozenset({"none", "ollama"})
 
 OLLAMA_PRESETS: dict[str, list[str]] = {
-    # Ollama library / community tags only (no GGUF). Tuned for SwarmUI prompt help:
-    # vision + RU/EN, share VRAM with diffusion on typical 24 GiB.
-    "recommended": ["huihui_ai/qwen2.5-vl-abliterated:7b"],  # ~6 GB, low refusal
-    "light": ["huihui_ai/qwen2.5-vl-abliterated:3b"],  # ~3 GB, tight VRAM
-    "stock": ["qwen2.5vl:7b"],  # official library vision
+    # Requirement: uncensored (abliterated) + solid Russian. Qwen2.5 family.
+    # Vision tags preferred for SwarmUI prompt help with images.
+    "recommended": ["huihui_ai/qwen2.5-vl-abliterated:7b"],  # ~6 GB
+    "light": ["huihui_ai/qwen2.5-vl-abliterated:3b"],  # ~3 GB
     "text": ["huihui_ai/qwen2.5-abliterate:7b"],  # ~5 GB, no vision
-    "big": ["qwen2.5vl:32b"],  # ~21 GB — llm-only or ≥40 GiB free
+    "big": ["huihui_ai/qwen2.5-vl-abliterated:32b"],  # ~21 GB
     "empty": [],
 }
 
 OLLAMA_PRESET_LABELS: dict[str, str] = {
-    "recommended": "Qwen2.5-VL 7B abliterate (картинки+RU, ~6GB)",
-    "light": "Qwen2.5-VL 3B abliterate (мало VRAM, ~3GB)",
-    "stock": "официальный qwen2.5vl:7b (library, больше отказов)",
-    "text": "Qwen2.5 7B abliterate text-only (~5GB)",
-    "big": "qwen2.5vl:32b (~21GB; llm-only / большой GPU)",
+    "recommended": "VL 7B abliterate — RU + картинки (~6GB)",
+    "light": "VL 3B abliterate — RU + картинки, мало VRAM (~3GB)",
+    "text": "7B abliterate text — RU, без vision (~5GB)",
+    "big": "VL 32B abliterate — RU + картинки (~21GB)",
     "empty": "только runtime, без pull",
     "keep": "не менять ollama-models.yaml",
 }