Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Qwen3-16B-A3B-UD-IQ1_S.gguf filter=lfs diff=lfs merge=lfs -text
+Qwen3-16B-A3B-UD-IQ1_M.gguf filter=lfs diff=lfs merge=lfs -text
+Qwen3-16B-A3B-UD-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
+Qwen3-16B-A3B-UD-Q2_K_XL.gguf filter=lfs diff=lfs merge=lfs -text
+Qwen3-16B-A3B-UD-Q4_K_XL.gguf filter=lfs diff=lfs merge=lfs -text

Qwen3-16B-A3B-UD-IQ1_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25c25a77953a78ae48c82065abe87bd6a08d7791901cc969d3b9235e182f8d14
+size 5251044032

Qwen3-16B-A3B-UD-IQ1_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ca56d774c49c9a6ad1861ffbd414858646590bcfa08dcd2fd8ec913f7db4c24
+size 4873638592

Qwen3-16B-A3B-UD-IQ2_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46744b84fc0498a1354827d848297f08e7ef2908b7e239a4b0adbea01fc1ae59
+size 5854958272

Qwen3-16B-A3B-UD-Q2_K_XL.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7f7a7db4bc621f5037ff75880c8cec2b96b202d403dc34ebfaa535b0bb045cd
+size 6342546112

Qwen3-16B-A3B-UD-Q4_K_XL.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6902a57785e5594a6b97162c6642d3d201743fb97d834423f4aae36b29e4891
+size 9310278336

README.md

@@ -1,16 +1,12 @@
 ---
 tags:
 - unsloth
-- qwen3
-- qwen
 license: apache-2.0
 base_model:
 - kalomaze/Qwen3-16B-A3B
 ---
 # Qwen3-16B-A3B
 
-Qwen3-16B-A3B is a rendition of Qwen3-30B-A3B by [kalomaze](https://huggingface.co/kalomaze/Qwen3-16B-A3B).
-
 A man-made horror beyond your comprehension.
 
 But no, seriously, this is my experiment to:
@@ -20,4 +16,4 @@ But no, seriously, this is my experiment to:
 It can still write semi-coherently without any additional training or distillation done on top of it from the original 30b MoE.
 The .txt files with the original measurements are provided in the repo along with the exported weights.
 
-Custom testing to measure the experts was done on a hacked version of vllm, and then I made a bespoke script to selectively export the weights according to the measurements.
+Custom testing to measure the experts was done on a hacked version of vllm, and then I made a bespoke script to selectively export the weights according to the measurements.

config.json

@@ -0,0 +1,39 @@
+{
+  "architectures": [
+    "Qwen3MoeForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "decoder_sparse_step": 1,
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "max_position_embeddings": 40960,
+  "max_window_layers": 48,
+  "mlp_only_layers": [],
+  "model_type": "qwen3_moe",
+  "moe_intermediate_size": 768,
+  "norm_topk_prob": true,
+  "num_attention_heads": 32,
+  "num_experts": 64,
+  "num_experts_per_tok": 8,
+  "num_hidden_layers": 48,
+  "num_key_value_heads": 4,
+  "output_router_logits": false,
+  "pad_token_id": 151654,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "router_aux_loss_coef": 0.001,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.0.dev0",
+  "unsloth_fixed": true,
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}