GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
HIMANSHUKUMARJHA
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
HIMANSHUKUMARJHA
Model Tree
Qwen/Qwen2.5-0.5B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Metric | Value |
|---|---|
| Eval loss (200 held-out rows) | 1.2073 |
Trained on 20,000 rows for 2,000 steps (~1.6 epochs). The earlier version used 2,000 rows for 200 steps, under-trained on 4% of the available data.
Caveat on the metric. Eval loss measures token prediction, not instruction-following quality.
python
from peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizerbase = "Qwen/Qwen2.5-0.5B-Instruct"tokenizer = AutoTokenizer.from_pretrained(base)model = AutoModelForCausalLM.from_pretrained(base, torch_dtype="bfloat16", device_map="auto")model = PeftModel.from_pretrained(model, "HIMANSHUKUMARJHA/autotune-qwen-lora")messages = [{'role': 'user', 'content': 'Give three tips for staying healthy.'}]inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True).to(model.device)inputs.pop("token_type_ids", None) # this architecture's generate() rejects itout = model.generate(**inputs, max_new_tokens=256)print(tokenizer.decode(out[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
Requires transformers>=4.51.
| Base model | Qwen/Qwen2.5-0.5B-Instruct |
| Dataset | yahma/alpaca-cleaned |
| Training rows | 20,000 |
| Steps | 2,000 (~1.6 epochs at effective batch 16) |
| LoRA rank / alpha | 32 / 64 |
| Target modules | attention + MLP projections |
| LR schedule | 0.0002 cosine, 3% warmup |
| Precision | bfloat16 |
| Hardware |
Reproduced with AutoTune:
bash
modal run finetune.py --profile alpaca
The dataset is shuffled with a fixed seed, then a 200-row holdout is taken before training. Rows the formatter cannot parse are dropped and exact duplicates removed, because these datasets contain repeats that would otherwise leak holdout examples into training.
This is a 0.5B parameter model. It is useful for general instruction following at small scale and for on-device or cost-sensitive settings, but it will not match a large general model. Outputs should be validated before use. The adapter inherits any bias present in the training dataset.
| 1x A10G |