Standardize model card (template rollout)

README.md

@@ -2,17 +2,18 @@
 license: other
 license_name: glm-5
 license_link: https://huggingface.co/zai-org/GLM-5.1/blob/main/LICENSE
-base_model: 0xSero/GLM-5.1-555B-A14B-REAP
+base_model:
+- 0xSero/GLM-5.1-555B
 tags:
-  - reap
-  - pruning
-  - moe
-  - expert-pruning
-  - glm
-  - gptq
-  - w4a16
-  - autoround
-  - vllm
+- autoround
+- expert-pruning
+- glm
+- gptq
+- moe
+- pruning
+- reap
+- vllm
+- w4a16
 library_name: transformers
 pipeline_tag: text-generation
 quantization_config:
@@ -22,9 +23,41 @@ quantization_config:
   sym: true
   desc_act: false
   checkpoint_format: gptq
+base_model_relation: quantized
 ---
 
-# GLM-5.1 — 25% Expert Pruned (REAP) — W4A16
+> [!TIP]
+> **[Support this work →](https://donate.sybilsolutions.ai)** · [X](https://x.com/0xsero) · [GitHub](https://github.com/0xsero) · [REAP paper](https://arxiv.org/abs/2510.13999) · [Cerebras REAP](https://huggingface.co/collections/cerebras/cerebras-reap)
+
+# GLM-5.1-555B-W4A16
+
+W4A16 quantization of [0xSero/GLM-5.1-555B](https://huggingface.co/0xSero/GLM-5.1-555B).
+
+## At a glance
+
+| | |
+|---|---|
+| Base model | [0xSero/GLM-5.1-555B](https://huggingface.co/0xSero/GLM-5.1-555B) |
+| Format | W4A16 |
+| Total params | **555B** |
+| Active / token | 14B |
+| Experts / layer | 192 |
+| Layers | 78 |
+| Hidden size | 6144 |
+| Context | 202,752 |
+| On-disk size | 297 GB |
+
+## Which variant should I pick?
+
+| Variant | Format | Link |
+|---|---|---|
+| `GLM-5.1-444B` | BF16 | [link](https://huggingface.co/0xSero/GLM-5.1-444B) |
+| `GLM-5.1-444B-GGUF` | GGUF | [link](https://huggingface.co/0xSero/GLM-5.1-444B-GGUF) |
+| `GLM-5.1-478B-NVFP4` | NVFP4 | [link](https://huggingface.co/0xSero/GLM-5.1-478B-NVFP4) |
+| `GLM-5.1-555B` | BF16 | [link](https://huggingface.co/0xSero/GLM-5.1-555B) |
+| `GLM-5.1-555B-GGUF` | GGUF | [link](https://huggingface.co/0xSero/GLM-5.1-555B-GGUF) |
+| `GLM-5.1-555B-NVFP4` | NVFP4 | [link](https://huggingface.co/0xSero/GLM-5.1-555B-NVFP4) |
+| `GLM-5.1-555B-W4A16` **(this)** | W4A16 | [link](https://huggingface.co/0xSero/GLM-5.1-555B-W4A16) |
 
 This is a **GPTQ 4-bit weight-quantized** variant of the 25% expert-pruned [`zai-org/GLM-5.1`](https://huggingface.co/zai-org/GLM-5.1) using [REAP](https://github.com/CerebrasResearch/reap) (Relative Expert Activation Pruning), produced with [AutoRound](https://github.com/intel/auto-round) for learned rounding optimization.
 
@@ -37,8 +70,8 @@ This is a **GPTQ 4-bit weight-quantized** variant of the 25% expert-pruned [`zai
 | Quantization | GPTQ W4A16 (int4 symmetric, group_size=128) |
 | Quantizer | auto-round 0.12.2 (200 iterations, SignSGD) |
 | Quantized size | **277 GB** (56 safetensor shards) |
-| BF16 source | [`0xSero/GLM-5.1-555B-A14B-REAP`](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP) |
-| GGUF variant | [`0xSero/GLM-5.1-555B-A14B-REAP-GGUF`](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GGUF) (325 GB, Q4_K_M) |
+| BF16 source | [`0xSero/GLM-5.1-555B`](https://huggingface.co/0xSero/GLM-5.1-555B) |
+| GGUF variant | [`0xSero/GLM-5.1-555B-GGUF`](https://huggingface.co/0xSero/GLM-5.1-555B-GGUF) (325 GB, Q4_K_M) |
 
 ## Benchmark Results (GGUF Q4_K_M, inference mode, temp=0.8)
 
@@ -61,7 +94,7 @@ The GPTQ W4A16 uses the same learned rounding method (AutoRound) as the GGUF Q4_
 from vllm import LLM, SamplingParams
 
 llm = LLM(
-    model="0xSero/GLM-5.1-555B-A14B-REAP-GPTQ-W4A16",
+    model="0xSero/GLM-5.1-555B-W4A16",
     tensor_parallel_size=4,    # 4× B200 or 8× A100
     max_model_len=8192,
     trust_remote_code=True,
@@ -75,7 +108,7 @@ outputs = llm.generate(["Hello, world!"], params)
 
 ```bash
 python -m sglang.launch_server \
-  --model-path 0xSero/GLM-5.1-555B-A14B-REAP-GPTQ-W4A16 \
+  --model-path 0xSero/GLM-5.1-555B-W4A16 \
   --tp 4 \
   --trust-remote-code
 ```
@@ -119,10 +152,10 @@ GPTQ packs 4-bit weights more efficiently with `group_size=128` symmetric quanti
 
 | Model | Prune % | Experts | Format | Size | Status |
 |-------|---------|---------|--------|------|--------|
-| [`0xSero/GLM-5.1-555B-A14B-REAP`](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP) | 25% | 192/256 | BF16 | 1.1T | Source checkpoint |
-| [`0xSero/GLM-5.1-555B-A14B-REAP-GGUF`](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GGUF) | 25% | 192/256 | GGUF Q4_K_M | 325G | llama.cpp serving |
+| [`0xSero/GLM-5.1-555B`](https://huggingface.co/0xSero/GLM-5.1-555B) | 25% | 192/256 | BF16 | 1.1T | Source checkpoint |
+| [`0xSero/GLM-5.1-555B-GGUF`](https://huggingface.co/0xSero/GLM-5.1-555B-GGUF) | 25% | 192/256 | GGUF Q4_K_M | 325G | llama.cpp serving |
 | **This model** | **25%** | **192/256** | **GPTQ W4A16** | **277G** | **vLLM/SGLang serving** |
-| [`0xSero/GLM-5.1-444B-A14B-REAP`](https://huggingface.co/0xSero/GLM-5.1-444B-A14B-REAP) | 40% | 154/256 | BF16 | 910G | Has repetition issues — use 25% |
+| [`0xSero/GLM-5.1-444B`](https://huggingface.co/0xSero/GLM-5.1-444B) | 40% | 154/256 | BF16 | 910G | Has repetition issues — use 25% |
 
 ## Support This Work
 
@@ -130,70 +163,16 @@ If you find these models useful, please consider supporting continued open-sourc
 
 **[donate.sybilsolutions.ai](https://donate.sybilsolutions.ai)**
 
-## Citation
-
-If you use this model, please cite the [REAP paper](https://github.com/CerebrasResearch/reap) and [AutoRound](https://github.com/intel/auto-round).
-
-## Sponsors
+## License & citation
+License inherited from the base model.
 
-Thank you for the kind sponsors, wouldn't be possible without them:
-
-- Nvidia
-- TNG Technology
-- Lambda
-- Prime Intellect
-- HotAisle
-
-<!-- GLM51_FAMILY_COMPAT_START -->
-## GLM-5.1 REAP Family — Hardware Compatibility
-
-All variants in this family are REAP-pruned ([2510.13999](https://arxiv.org/abs/2510.13999)) descendants of [zai-org/GLM-5.1](https://huggingface.co/zai-org/GLM-5.1) (original: 744B params, 256 experts/MoE layer, 40B activated/token). Pick a variant based on your GPU architecture and available VRAM.
-
-### Quick picker
-
-| You have | Use |
-|---|---|
-| 8× H100/H200 80GB (Hopper, sm_90) | **[GLM-5.1-555B-A14B-REAP-GPTQ-W4A16](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GPTQ-W4A16)** or **[GLM-5.1-555B-A14B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-NVFP4)** (NVFP4 on Hopper via `modelopt_fp4` + triton path) |
-| 4× RTX PRO 6000 Blackwell Workstation 96GB (sm_120) | **[GLM-5.1-478B-A42B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-478B-A42B-REAP-NVFP4)** (further-pruned 160-expert, 200k ctx) — this is the Blackwell Workstation reference config |
-| 4× B200 180GB (sm_100) | **[GLM-5.1-478B-A42B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-478B-A42B-REAP-NVFP4)** or **[GLM-5.1-555B-A14B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-NVFP4)** |
-| 8× B200 / Blackwell datacenter | **[GLM-5.1-555B-A14B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-NVFP4)** (192-expert, upstream's reference config with flashinfer + b12x backends) |
-| 8× A100 80GB (Ampere, sm_80) | **[GLM-5.1-444B-A14B-REAP](https://huggingface.co/0xSero/GLM-5.1-444B-A14B-REAP)** (BF16) or **[-GPTQ-W4A16](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GPTQ-W4A16)** |
-| CPU / Apple Silicon / consumer GPU with llama.cpp | **[GLM-5.1-555B-A14B-REAP-GGUF](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GGUF)** or **[GLM-5.1-444B-A14B-REAP-GGUF](https://huggingface.co/0xSero/GLM-5.1-444B-A14B-REAP-GGUF)** |
-
-### Full family
-
-| Variant | Format | Size | Experts/layer | Activated/token | Min VRAM (TP) | Inference engine | Best on |
-|---|---|---|---|---|---|---|---|
-| [GLM-5.1-555B-A14B-REAP](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP) | BF16 | ~1125 GB | 192 | ~14B | 8× 141 GB (H200) | sglang / vllm | Hopper |
-| [GLM-5.1-444B-A14B-REAP](https://huggingface.co/0xSero/GLM-5.1-444B-A14B-REAP) | BF16 | ~910 GB | 154 | ~14B | 8× 114 GB | sglang / vllm | Ampere / Hopper |
-| [GLM-5.1-555B-A14B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-NVFP4) | NVFP4 (4-bit) | ~320 GB | 192 | ~14B | 4× 80 GB (B200), 8× 48 GB | sglang `--quantization modelopt_fp4` | Blackwell (native); Hopper (triton path) |
-| [GLM-5.1-478B-A42B-REAP-NVFP4](https://huggingface.co/0xSero/GLM-5.1-478B-A42B-REAP-NVFP4) | NVFP4 (4-bit) | ~285 GB | 160 | ~42B | 4× 80 GB Blackwell | sglang `--quantization modelopt_fp4` | 4× RTX PRO 6000 Blackwell @ 200k ctx |
-| [GLM-5.1-555B-A14B-REAP-GPTQ-W4A16](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GPTQ-W4A16) | GPTQ W4A16 | ~297 GB | 192 | ~14B | 4× 80 GB | vllm / sglang `--quantization gptq_marlin` | Hopper (best), works on Ampere |
-| [GLM-5.1-555B-A14B-REAP-GGUF](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-GGUF) | GGUF (Q2–Q8) | ~348 GB | 192 | ~14B | Varies by quant | llama.cpp | CPU / Apple / consumer CUDA |
-| [GLM-5.1-444B-A14B-REAP-GGUF](https://huggingface.co/0xSero/GLM-5.1-444B-A14B-REAP-GGUF) | GGUF (Q2–Q8) | ~283 GB | 154 | ~14B | Varies by quant | llama.cpp | CPU / Apple / consumer CUDA |
-
-### Notes
-
-- **NVFP4 on Hopper (H100/H200)**: supported from sglang 25.10 / 0.5.10+ ([NVIDIA SGLang release notes](https://docs.nvidia.com/deeplearning/frameworks/sglang-release-notes/)); native Blackwell tensor-core FP4 still gives better throughput.
-- **NVFP4 on B200 / Blackwell datacenter (sm_100)**: use flashinfer attention + `b12x` or flashinfer MoE backends — this is the recipe in the original [555B-A14B-REAP-NVFP4 card](https://huggingface.co/0xSero/GLM-5.1-555B-A14B-REAP-NVFP4).
-- **NVFP4 on Blackwell Workstation (sm_120)**: use `--attention-backend triton` (not flashinfer — PCIe P2P atomics unavailable on the consumer board), `--moe-runner-backend cutlass`, `--fp4-gemm-backend flashinfer_cudnn`. See the [GLM-5.1-478B-A42B-REAP-NVFP4 card](https://huggingface.co/0xSero/GLM-5.1-478B-A42B-REAP-NVFP4) for the full 200k-ctx replication guide.
-- **GPTQ-W4A16 vs NVFP4**: same bit depth, different hardware path. NVFP4 has native Blackwell support and per-16 fp8 scales; GPTQ is group-quantized int4 with broader engine support.
-- **REAP expert count variants (555B/444B)**: different expert-retention ratios from the same base; 555B keeps more experts (higher quality ceiling), 444B trades quality for 20% less VRAM.
-- **Why NVFP4-478B-A42B-REAP is different**: it's double-pruned (256 → 192 → 160 experts), optimized for a specific Blackwell Workstation 4×96GB target at 200k context. The A42B suffix reflects measured activated params/token on the 160-expert MoE, not the REAP branding convention of the sibling variants.
-
-### Pointer to active inference recipe
-
-See [GLM-5.1-478B-A42B-REAP-NVFP4 README](https://huggingface.co/0xSero/GLM-5.1-478B-A42B-REAP-NVFP4) for the full Blackwell Workstation replication guide (exact software pins, NSA patch, launch flags, measured 200k-ctx perf, sampling recommendations). Most of the sglang flags carry over to other NVFP4 variants on other hardware.
-
-### Citation
-```
+```bibtex
 @misc{lasby2025reap,
-  title={REAP the Experts: Why Pruning Prevails for One-Shot MoE compression},
-  author={Mike Lasby and Ivan Lazarevich and Nish Sinnadurai and Sean Lie and Yani Ioannou and Vithursan Thangarasa},
-  year={2025},
-  eprint={2510.13999},
-  archivePrefix={arXiv},
+  title  = {REAP the Experts: Why Pruning Prevails for One-Shot MoE Compression},
+  author = {Mike Lasby and Ivan Lazarevich and Nish Sinnadurai and Sean Lie and Yani Ioannou and Vithursan Thangarasa},
+  year   = {2025}, eprint = {2510.13999}, archivePrefix = {arXiv}
 }
 ```
 
-<!-- GLM51_FAMILY_COMPAT_END -->
+## Sponsors
+Made possible by **NVIDIA · TNG Technology · Lambda · Prime Intellect · Hot Aisle**.