---
language:
- ru
license: other
license_name: gemma
license_link: https://ai.google.dev/gemma/terms
base_model: google/gemma-4-E4B-it
pipeline_tag: text-generation
tags:
- gemma4
- russian
- colloquial
- style-transfer
- merged
- vllm
library_name: transformers
datasets:
- pavelfedortsov/russian-colloquial-sft-50k
---

# gemma4-e4b-colloquial-ru-merged

*English:* Full-weight **Gemma 4 E4B** checkpoint with colloquial Russian LoRA **merged in** for vLLM / [RunPod Serverless](https://www.runpod.io/serverless). No PEFT at inference time.

## What this model does

Rewrites **formal Russian** into **casual chat-style Russian** (Telegram-like), **without profanity**, while keeping facts, names, numbers, and paragraph structure.

**Not** a general chat model — use the instruction prefix from training (see below).

## Model lineage

| Stage | Artifact |
|-------|----------|
| Base | [google/gemma-4-E4B-it](https://huggingface.co/google/gemma-4-E4B-it) |
| LoRA (SFT) | [pavelfedortsov/gemma4-e4b-lora-colloquial-ru](https://huggingface.co/pavelfedortsov/gemma4-e4b-lora-colloquial-ru) |
| **This repo** | LoRA merged into base + vLLM fixes (`k_norm`, processor configs) |

Merge was done with `peft.merge_and_unload()`; missing `language_model` **k_norm** weights for layers 24–41 were copied from the base checkpoint (required for vLLM).

## Training data

- **50,000** SFT pairs, mat-free colloquial style  
- Hub dataset: [pavelfedortsov/russian-colloquial-sft-50k](https://huggingface.co/datasets/pavelfedortsov/russian-colloquial-sft-50k)  
- Built from [kurumikz/telegram-corpus-russian-kazakh](https://huggingface.co/datasets/kurumikz/telegram-corpus-russian-kazakh) + Gemini pair generation (see dataset card)

**User prompt template (training & inference):**

```
Перепиши простым разговорным русским, как в переписке. Без мата и грубости. Сохрани смысл:
<формальный текст>
```

## Training configuration (LoRA → merge)

Config file (also in `card_assets/train_colloquial_e4b_gpu.yaml`):

| Parameter | Value |
|-----------|--------|
| Base model | `google/gemma-4-E4B-it` |
| Method | LoRA on language tower (`model.language_model.*`) |
| LoRA rank / alpha | **32 / 64** |
| Target modules | `q,k,v,o` + MLP (`gate, up, down`) |
| Dataset | 50k × 1 repeat |
| Epochs | **2** (12,500 optimizer steps) |
| Seq length | 512 |
| Batch | 1 × grad accum **8** (effective 8) |
| LR | 1e-4, cosine, warmup 3% |
| Precision | bf16, gradient checkpointing |
| Loss | assistant-only |
| Hardware | RunPod **A100 80GB** |

## Training metrics (LoRA run)

![Training curves](card_assets/training_curves.png)

| Metric | Start (step ~25) | End (step 12,500) | Best |
|--------|------------------|-------------------|------|
| `loss` | ~3.42 | ~0.81 | **~0.67** |
| `mean_token_accuracy` | ~0.63 | ~0.82 | **~0.84** |

Checkpoints saved every 1000 steps under the LoRA adapter repo.

## Inference

### RunPod Serverless (vLLM)

```env
MODEL_NAME=pavelfedortsov/gemma4-e4b-colloquial-ru-merged
HF_TOKEN=<your_token>
TRUST_REMOTE_CODE=true
DTYPE=bfloat16
MAX_MODEL_LEN=4096
GPU_MEMORY_UTILIZATION=0.90
ENFORCE_EAGER=true
ENABLE_LORA=false
LANGUAGE_MODEL_ONLY=true
LIMIT_MM_PER_PROMPT={"image":0,"audio":0,"video":0}
```

Recommended GPU: **≥40 GB** VRAM (merged ~32 GB weights in bf16).

### Transformers (local)

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "pavelfedortsov/gemma4-e4b-colloquial-ru-merged"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

formal = "Сегодня на совещании обсуждали внедрение новой версии API."
user = (
    "Перепиши простым разговорным русским, как в переписке. "
    "Без мата и грубости. Сохрани смысл:\n"
    f"{formal}"
)
messages = [{"role": "user", "content": user}]
prompt = tokenizer.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=256, do_sample=True, temperature=0.7, top_p=0.9)
print(tokenizer.decode(out[0, inputs["input_ids"].shape[1]:], skip_special_tokens=True))
```

### OpenAI-compatible API (RunPod / vLLM)

```bash
curl "$RUNPOD_URL/v1/chat/completions" \
  -H "Authorization: Bearer $RUNPOD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pavelfedortsov/gemma4-e4b-colloquial-ru-merged",
    "messages": [{
      "role": "user",
      "content": "Перепиши простым разговорным русским, как в переписке. Без мата и грубости. Сохрани смысл:\nВаш формальный текст."
    }],
    "max_tokens": 512,
    "temperature": 0.7
  }'
```

## Limitations

- [Gemma license](https://ai.google.dev/gemma/terms) applies to the base architecture and weights.
- Quality varies on long news-style text; model may shorten or paraphrase aggressively.
- Not safety-tuned for production without your own evaluation.
- Merged vs LoRA inference can differ slightly in style.

## Related repos

| Resource | Link |
|----------|------|
| LoRA adapter | https://huggingface.co/pavelfedortsov/gemma4-e4b-lora-colloquial-ru |
| Dataset (50k) | https://huggingface.co/datasets/pavelfedortsov/russian-colloquial-sft-50k |
| Base model | https://huggingface.co/google/gemma-4-E4B-it |