model-00006.safetensors 11 GBmodel-00012.safetensors 11 GBmodel-00008.safetensors 11 GBmodel-00005.safetensors 11 GBmodel-00013.safetensors 11 GBmodel-00002.safetensors 11 GBmodel-00007.safetensors 11 GBmodel-00011.safetensors 11 GBmodel-00009.safetensors 10 GBmodel-00003.safetensors 10 GBmodel-00004.safetensors 10 GBmodel-00010.safetensors 10 GBmodel-00001.safetensors 10 GBmodel-00014.safetensors 2.3 GBMLX (Apple Silicon) build of GLM-5.3-Flash — 320B-A18B hybrid of 34 Kimi-Delta linear-attention layers and 11 DeepSeek-sparse-attention (NoPE MLA + lightning indexer) layers with manifold-constrained hyper-connections — quantized to 4-bit experts / 8-bit everything else.
These files are modified: converted from the upstream bfloat16 release (GLM-5.3-Flash-BF16) to MLX and quantized; the architecture is unchanged. The multi-token-prediction layer (layer 45) is not included. The vision tower is carried in bfloat16.
glm5_next landed in mlx-vlm main on 2026-08-26 (no
release carries it yet). Validating that port against transformers 5.16 at tiny scale found two
numerical bugs and two epsilon mismatches, which the runtime in
https://github.com/PipeNetwork/glm53-flash-mlx fixes; parity is 1e-6 end to end, exact on cached decode.
| what | reference | mlx-vlm main |
effect |
|---|---|---|---|
swiglu_limit |
gate clamped at 10, up at ±10, in every text MLP | no clamp anywhere in the text stack | formula mismatch on all 45 FFN blocks |
mHC base/scale dtype |
float32 | converter casts to bf16; the Metal kernel then reads base as float4 |
comb mixing matrix off by ~0.5 on every layer of a converted checkpoint |
| MLA low-rank norm eps | rms_norm_eps = 1e-5 |
1e-6 | small |
| indexer LayerNorm eps | 1e-6 | 1e-5 | small |
This checkpoint keeps the mHC arrays and KDA decay parameters in float32 as stored, so it is safe in either runtime; the clamp is a compute-path fix and needs the patched runtime:
git clone https://github.com/PipeNetwork/glm53-flash-mlx && cd glm53-flash-mlx && pip install -r requirements.txt
python scripts/smoke_generate.py /path/to/GLM-5.3-Flash-REAP25-MLX-mixed-4_8bit
from glm53_flash_mlx.load import load
model, processor = load("/path/to/GLM-5.3-Flash-REAP25-MLX-mixed-4_8bit")
This build keeps 216 of 288 routed experts per MoE layer (25% pruned; the 3 dense layers, attention, shared experts, router and vision tower are untouched), chosen by REAP saliency — mean router_weight × ‖expert_output‖ over 65,536 calibration tokens (wikitext-2 train, ten languages of Wikipedia and code; zero 32-gram overlap with the eval set), collected by running the full 8-bit build. Kept experts carry 86.9% of saliency mass on average; two disjoint halves of the calibration set pick the same kept set 93.3% of the time. The pruning was applied to the already-quantized build (equivalent to pruning bf16 and requantizing). Saliency retention is not a quality measure — the perplexity below is.
139.1 GB on disk (bfloat16 upstream: 642.7 GB).
| group | share of parameters | this build |
|---|---|---|
routed experts (switch_mlp, 42 layers × 288) |
304B (97%) | 4-bit, group 64 |
KDA and MLA projections, shared experts, dense MLPs, embeddings, lm_head |
~9B | 8-bit, group 64 |
| lightning-indexer projections | 0.06B | 8-bit, group 64 |
MoE router + correction bias, mHC arrays (fp32), KDA A_log/dt_bias (fp32), convolutions, norms |
— | as stored |
| vision tower | 0.56B | bfloat16 |
Perplexity on wikitext-2 (test), 288,627 tokens in 141 windows of 2048, every build scored on identical windows through this runtime. The 643 GB bfloat16 model does not fit a 512 GB machine, so the 8-bit build is the anchor (on every model we have measured, 8-bit has been statistically indistinguishable from bfloat16). Per-window NLL differences against 8-bit, bootstrapped over one shared index set (20,000 resamples):
| build | size | perplexity | ΔNLL/token vs 8-bit [95% CI] | windows worse |
|---|---|---|---|---|
| 8bit | 334.1 GB | 3.4607 | — | — |
| 6bit | 255.9 GB | 3.4646 | +0.0011 [−0.0017, +0.0038] | 89/141 |
| mixed-4_8bit | 181.9 GB | 3.5705 | +0.0312 [+0.0271, +0.0355] | 131/141 |
| 4bit | 177.6 GB | 3.7549 | +0.0816 [+0.0755, +0.0879] | 140/141 |
| REAP25-mixed-4_8bit | 139.1 GB | 4.2249 | +0.1995 [+0.1657, +0.2377] | 139/141 |
| REAP37-mixed-4_8bit | 118.3 GB | 4.8752 | +0.3427 [+0.2968, +0.3929] | 141/141 |
| REAP50-mixed-4_8bit | 96.3 GB | 6.0757 | +0.5628 [+0.5071, +0.6219] | 141/141 |
| REAP25-4bit | — | 4.4361 | +0.2483 [+0.2135, +0.2873] | 141/141 |
| REAP37-4bit | — | 5.1057 | +0.3889 [+0.3424, +0.4393] | 141/141 |
| REAP50-4bit | — | 6.3840 | +0.6123 [+0.5552, +0.6722] | 141/141 |
Read the interval, not the point estimate; "windows worse" counts how many of the 141 windows the build lost outright.
Against the 8-bit anchor: 6-bit +0.1%, mixed 4/8-bit +3.2%, uniform 4-bit +8.5%. Routed experts are 97% of the parameters; the mixed build keeps the other ~9B (KDA and MLA projections, shared experts, dense layers, embeddings) at 8-bit for 4.4 GB more than uniform 4-bit.
Greedy generation (a collapse detector, not a ranking) is coherent on every published build.
MIT, as the upstream model. Port code: https://github.com/PipeNetwork/glm53-flash-mlx.
有文件？提交种子。
提交种子