Upload folder using huggingface_hub

README.md

@@ -0,0 +1,77 @@
+---
+language: zh
+tags:
+- qwen3
+- lora
+- fine-tuning
+- translation
+base_model: Qwen/Qwen3-8B
+---
+
+# Qwen3-8B-Translator-LoRA
+
+This model is a fine-tuned version of `Qwen/Qwen3-8B` using LoRA for English to Chinese translation, specifically tailored for audio product terminology.
+
+## Fine-tuning Details
+
+- **Fine-tuning Method:** LoRA (Low-Rank Adaptation)
+- **Dataset:** Custom parallel corpus for audio products (English-Chinese)
+- **Framework:** PyTorch, Hugging Face Transformers, TRL, PEFT, Optimum TPU
+- **Hardware:** Google Cloud TPU v3-8
+
+## Training Procedure
+
+The model was trained using the `SFTTrainer` from the TRL library.
+
+### Training Hyperparameters
+
+- `max_seq_length`: 768
+- `per_device_train_batch_size`: 32
+- `per_device_eval_batch_size`: 32
+- `num_train_epochs`: 10
+- `eval_strategy`: "steps"
+- `eval_steps`: 10
+- `learning_rate`: 2e-5
+- `lr_scheduler_type`: "cosine"
+- `warmup_ratio`: 0.1
+- `weight_decay`: 0.01
+- `optim`: "adamw_torch_xla"
+
+### LoRA Configuration
+
+- `r`: 128
+- `lora_alpha`: 256
+- `lora_dropout`: 0.05
+- `bias`: "none"
+- `target_modules`: ["q_proj", "v_proj", "gate_proj", "down_proj"]
+- `modules_to_save`: ["lm_head", "embed_tokens"]
+
+## Training Results
+
+| Step | Training Loss | Validation Loss |
+|------|---------------|-----------------|
+| 10 | 0.844400 | 0.635387 |
+| 20 | 0.486000 | 0.407656 |
+| 30 | 0.439900 | 0.381002 |
+| 40 | 0.391100 | 0.365226 |
+| 50 | 0.370300 | 0.352978 |
+| 60 | 0.307100 | 0.345395 |
+| 70 | 0.368900 | 0.340513 |
+| 80 | 0.306000 | 0.335354 |
+| 90 | 0.273900 | 0.333215 |
+| 100 | 0.272400 | 0.334439 |
+| 110 | 0.256300 | 0.331390 |
+| 120 | 0.226100 | 0.334290 |
+| 130 | 0.246800 | 0.338176 |
+| 140 | 0.230500 | 0.339353 |
+
+## Intended Use
+
+This model is intended for translating English text related to audio products into Chinese. It can be used by professionals in the audio industry, technical writers, or anyone needing to translate such content.
+
+## Limitations and Bias
+
+- The model's performance is best on text similar to the data it was trained on (audio product domain).
+- It may not generalize well to other domains or highly colloquial language.
+- As with any language model, there's a potential for biases present in the training data to be reflected in the output.
+    

adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/kaggle/input/qwen-3/transformers/8b/1",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": [
+    "lm_head",
+    "embed_tokens"
+  ],
+  "peft_type": "LORA",
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "q_proj",
+    "down_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48d382385148dbdc194f78e0a465c825f9455ed8ab63844cb73097fa2107075e
+size 3338705488