GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
ndtran0101
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
ndtran0101
Model Tree
Qwen/Qwen3-Embedding-8B
Input Modalities
Output Modalities
Supported Functionality
| metric | value | frozen-embedding baseline |
|---|---|---|
| MAE | 0.581 | 0.925 |
| within ±0.5 band | 71.8% | 47.8% |
| within ±1.0 band | 89.9% | — |
| σ-interval coverage (±1σ / ±2σ) | 0.73 / 0.96 | — |
Strongest in the dense middle-to-upper region (bands 4.5–5.0, 7.0–8.0: MAE 0.43–0.54). Weakest at the extremes (band 9: 0.87) — the residual there is dominated by label noise in the training source (independent expert graders themselves disagree ~0.4 bands), not model capacity.
text
essay ─▶ Qwen3-Embedding-8B + this LoRA ─▶ 4096-d last-token embedding ─▶ head ─▶ (μ, σ) ─▶ band (last-token pool, NO L2 normalization)q/k/v/o + gate/up/down projections.head_np.npz): Linear(4096→512) → GELU → Linear(512→2) → (μ, log σ²). Band =
round-to-0.5(μ); σ = exp(½·logσ²)·T with temperature T=1.076 (see meta.json).python
import numpy as np, torch, torch.nn as nn, math, jsonfrom transformers import AutoModel, AutoTokenizerfrom peft import PeftModelfrom huggingface_hub import hf_hub_download REPO = "ndtran0101/qwen3-embedding-8b-ielts-band-lora"tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-Embedding-8B"); tok.padding_side = "left"base = AutoModel.from_pretrained("Qwen/Qwen3-Embedding-8B", torch_dtype=torch.bfloat16)model = PeftModel.from_pretrained(base, REPO).eval().cuda() h = np.load(hf_hub_download(REPO, "head_np.npz")); meta = json.load(open(hf_hub_download(REPO, "meta.json")))def gelu(x): return 0.5*x*(1+np.vectorize(math.erf)(x/math.sqrt(2))) def score(essay, question=""): q = question.strip(); body = essay.strip() text = f"Instruct: {meta['instruct']}\nQuery: {q}\n\n{body}" + tok.eos_token enc = tok([text], return_tensors="pt", truncation=True, max_length=768).to("cuda") with torch.no_grad(), torch.autocast("cuda", dtype=torch.bfloat16): emb = model(**enc).last_hidden_state[:, -1, :].float().cpu().numpy()[0] # raw, un-normalized o = gelu(emb @ h["0_weight"].T + h["0_bias"]) @ h["3_weight"].T + h["3_bias"] mu = float(o[0]); sigma = math.exp(0.5*float(np.clip(o[1], -6, 4)))*meta["temperature"] return {"band": round(min(max(mu,0),9)*2)/2, "mu": round(mu,3), "sigma": round(sigma,3)} print(score("Living in a city has many advantages ...", question="Some people prefer city life ..."))Serve the LoRA-adapted encoder as an embedding endpoint (last-token pool, no normalize), then
run the head client-side with the bundled score_band.py:
bash
docker run --restart unless-stopped -d --name qwen-embeddings-8b \ --gpus '"device=0"' -v ~/.cache:/root/.cache -p 8000:8000 --ipc host \ vllm/vllm-openai:nightly \ Qwen/Qwen3-Embedding-8B \ --served-model-name qwen/qwen3-embedding-8b \ --runner pooling \ --pooler-config '{"pooling_type":"LAST","use_activation":false}' \ --enable-lora \ --lora-modules qwen/qwen3-embedding-8b:ielts-band=ndtran0101/qwen3-embedding-8b-ielts-band-lora \ --max-lora-rank 16Then embed with model="qwen/qwen3-embedding-8b:ielts-band" and feed the vector to the head. See score_band.py.
| file | purpose |
|---|---|
adapter_config.json, adapter_model.safetensors | the LoRA adapter (PEFT) |
head_np.npz | regression head weights (numpy) |
meta.json | serving config (instruction, pooling, temperature, dims) |
score_band.py | reference client (embed → head → band) |
Fine-tuned end-to-end (LoRA on the encoder + head) on
ndtran0101/writing9-ielts-essays
(114k train). Objective: Gaussian β-NLL (heteroscedastic mean–variance regression; Nix & Weigend
1994, Kendall & Gal 2017) with inverse-frequency class weighting and post-hoc temperature
calibration of σ. Config: lr 2e-4, class-weight power 1.0, β 0.5, last-token pooling, 2 epochs,
plus 2,500 LLM-polished band-9 examples for tail augmentation.
Trained against crowd-sourced band labels that are noisy at the extremes; treat single-band precision for bands ≤4.5 and ≥8.5 with caution and prefer the ±σ range. English only.
import numpy as np, torch, torch.nn as nn, math, jsonfrom transformers import AutoModel, AutoTokenizerfrom peft import PeftModelfrom huggingface_hub import hf_hub_download REPO = "ndtran0101/qwen3-embedding-8b-ielts-band-lora"tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-Embedding-8B"); tok.padding_side = "left"base = AutoModel.from_pretrained("Qwen/Qwen3-Embedding-8B", torch_dtype=torch.bfloat16)model = PeftModel.from_pretrained(base, REPO).eval().cuda() h = np.load(hf_hub_download(REPO, "head_np.npz")); meta = json.load(open(hf_hub_download(REPO, "meta.json")))def gelu(x): return 0.5*x*(1+np.vectorize(math.erf)(x/math.sqrt(2))) def score(essay, question=""): q = question.strip(); body = essay.strip() text = f"Instruct: {meta['instruct']}\nQuery: {q}\n\n{body}" + tok.eos_token enc = tok([text], return_tensors="pt", truncation=True, max_length=768).to("cuda") with torch.no_grad(), torch.autocast("cuda", dtype=torch.bfloat16): emb = model(**enc).last_hidden_state[:, -1, :].float().cpu().numpy()[0] # raw, un-normalized o = gelu(emb @ h["0_weight"].T + h["0_bias"]) @ h["3_weight"].T + h["3_bias"] mu = float(o[0]); sigma = math.exp(0.5*float(np.clip(o[1], -6, 4)))*meta["temperature"] return {"band": round(min(max(mu,0),9)*2)/2, "mu": round(mu,3), "sigma": round(sigma,3)} print(score("Living in a city has many advantages ...", question="Some people prefer city life ..."))docker run --restart unless-stopped -d --name qwen-embeddings-8b \ --gpus '"device=0"' -v ~/.cache:/root/.cache -p 8000:8000 --ipc host \ vllm/vllm-openai:nightly \ Qwen/Qwen3-Embedding-8B \ --served-model-name qwen/qwen3-embedding-8b \ --runner pooling \ --pooler-config '{"pooling_type":"LAST","use_activation":false}' \ --enable-lora \ --lora-modules qwen/qwen3-embedding-8b:ielts-band=ndtran0101/qwen3-embedding-8b-ielts-band-lora \ --max-lora-rank 16