Instructions to use 2wild4tv/GLM-5.3-Int4-Int8Mix with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use 2wild4tv/GLM-5.3-Int4-Int8Mix with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="2wild4tv/GLM-5.3-Int4-Int8Mix")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("2wild4tv/GLM-5.3-Int4-Int8Mix")
model = AutoModelForCausalLM.from_pretrained("2wild4tv/GLM-5.3-Int4-Int8Mix", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use 2wild4tv/GLM-5.3-Int4-Int8Mix with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "2wild4tv/GLM-5.3-Int4-Int8Mix"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "2wild4tv/GLM-5.3-Int4-Int8Mix",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/2wild4tv/GLM-5.3-Int4-Int8Mix

SGLang

How to use 2wild4tv/GLM-5.3-Int4-Int8Mix with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "2wild4tv/GLM-5.3-Int4-Int8Mix" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "2wild4tv/GLM-5.3-Int4-Int8Mix",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "2wild4tv/GLM-5.3-Int4-Int8Mix" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "2wild4tv/GLM-5.3-Int4-Int8Mix",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use 2wild4tv/GLM-5.3-Int4-Int8Mix with Docker Model Runner:
```
docker model run hf.co/2wild4tv/GLM-5.3-Int4-Int8Mix
```

2wild4tv commited on about 4 hours ago

Commit

5fa383e

verified ·

1 Parent(s): 02ba89d

Card: verified evidence (structural + byte-for-byte QuantTrio match + coherence), honest no-parity caveat, sm12x overlay serving note

Browse files

Files changed (1) hide show

README.md +13 -7

README.md CHANGED Viewed

@@ -60,18 +60,24 @@ vllm serve <path-to-this-model> \
   --max-model-len 131072   # raise per your KV pool
 ```
 - **Speculative decode:** pairs with **[incoai/GLM-5.3-DFlash2](https://huggingface.co/incoai/GLM-5.3-DFlash2)** (a 4.9 GB DFlash2 drafter) at ~zero extra KV pool — the flagship stack. Native MTP (layer 78) is preserved as a fallback.
 - **Long context:** add **Decode Context Parallel** (`--decode-context-parallel-size 4`) to split the KV across all four nodes for a ~600K single-request context (the tradeoff is a small cross-node cost per decode step).
-## Benchmarks
-Measured on 4x DGX Spark (GB10, sm121, aarch64) over a RoCE fabric. **TBD — filling in after the serve-test.**
-| config | decode tok/s | context | KV pool | date |
-|---|---|---|---|---|
-| Int4-Int8Mix, fp8 KV, TP4 | TBD | TBD | TBD | TBD |
-| + DFlash2 | TBD | TBD | TBD | TBD |
-| + DCP4 | TBD | ~600K | TBD | TBD |
 ## Credits (what's whose)

   --max-model-len 131072   # raise per your KV pool
 ```
+- **⚠️ GB10 / sm121 needs the sm12x kernel overlays** (the sparse-MLA + fp8-KV path is not in stock vLLM for this arch). Get them from the sibling recipe repo: **[tonyd2wild/GLM-5.2-QuantTrio-200K-4x-DGX-Spark](https://github.com/tonyd2wild/GLM-5.2-QuantTrio-200K-4x-DGX-Spark--36tok-s)**. On standard CUDA (H100/A100 etc.), `--quantization compressed-tensors` is all you need.
+- `kv_cache_scheme: None` in the config — KV precision is a **serve-time** choice (`--kv-cache-dtype`), not baked into the weights.
 - **Speculative decode:** pairs with **[incoai/GLM-5.3-DFlash2](https://huggingface.co/incoai/GLM-5.3-DFlash2)** (a 4.9 GB DFlash2 drafter) at ~zero extra KV pool — the flagship stack. Native MTP (layer 78) is preserved as a fallback.
 - **Long context:** add **Decode Context Parallel** (`--decode-context-parallel-size 4`) to split the KV across all four nodes for a ~600K single-request context (the tradeoff is a small cross-node cost per decode step).
+## Verification (what's proven, and what isn't)
+Produced by a shard-streaming RTN quantizer (reads one BF16 shard at a time, quantizes with `compressed_tensors`' own `calculate_qparams` / `quantize` / `pack_to_int32`, writes a 1:1 output shard; peak RAM ~10 GiB, runtime **28.2 min**). NOT `llmcompressor.oneshot` — accelerate disk-offload would need ~1.4 TB of scratch this fleet doesn't have.
+**Proven:**
+- **Structural (fail-closed):** 59,585 source tensors → 58,992 quantized modules → 177,569 output tensors (exact). Group counts **57,600 / 616 / 776** match the pre-quant dry-run exactly (75 layers × 256 experts × 3 = 57,600 ✓). Sacred modules (layer 0, `mlp.gate`, indexer, `lm_head`, MTP norms) all plain BF16, zero packed leaks. 282 shards, none missing.
+- **Layout matches QuantTrio byte-for-byte** (verified via HTTP range-reads of their published safetensors headers): expert `down_proj.weight_packed` I32 `[6144,256]`, `weight_scale` BF16 `[6144,16]`, MTP scale BF16 `[6144,1]` — all identical.
+- **Numerical round-trip:** int8 group/128 ≈ 0.70% rel error, int4 ≈ 12% (normal for 16 levels), full range used.
+- **Serves live:** TP4 across 4× DGX Spark (GB10/sm121), vLLM 0.23.1rc1, weights 95.53 GiB/rank, fp8_ds_mla KV, 202,944-token KV pool @ 200K ctx, health 200. Passed the bat-and-ball coherence trap ($0.05 with correct algebra) — a mis-quantized router or indexer cannot do that.
+**NOT yet done (honest):**
+- **No 69-scenario eval.** Coherence is proven; quality-vs-fp8 parity is NOT measured. Do not assume parity.
+- **Speed:** not benchmarked representatively. Any current number is thinking-ON, no spec-decode, no CUDA graphs = a FLOOR, not the model's real speed. Proper DFlash2 + graphs numbers will be added when measured.
 ## Credits (what's whose)