Upload folder using huggingface_hub

.gitattributes

@@ -63,3 +63,4 @@ minicpm5/gguf/model-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
 tinyllama-chat/gguf/model-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
 gemma4-random-model/gguf-q4_0/gemma4-random-model-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
 qwen3-random-model/gguf-q4_0/qwen3-random-model-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
+smollm3-random-model/gguf-q4_0/smollm3-random-model-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text

ARCHITECTURE_RANDOM_MODELS_REPORT.json

@@ -0,0 +1,57 @@
+{
+  "cases": [
+    {
+      "architecture": "gemma4",
+      "case": "gemma4-random-model",
+      "eog_token_ids": [
+        1
+      ],
+      "gguf_quantization": "Q4_0",
+      "gguf_sha256": "2b631587994acd127d3e6fcc67552c707d6fdba22beada963597387b9e814881",
+      "hf_dtype": "bfloat16",
+      "hf_sha256": "18c8b09759c831fbc9d6a8caf7df8e9d395508d4aec09336167f675527b3a360",
+      "parameter_count": 1519168,
+      "tensor_type_histogram": {
+        "F32": 44,
+        "Q4_0": 42
+      },
+      "validation": "passed"
+    },
+    {
+      "architecture": "qwen3",
+      "case": "qwen3-random-model",
+      "eog_token_ids": [
+        1
+      ],
+      "gguf_quantization": "Q4_0",
+      "gguf_sha256": "6af2e0d230a7919c6b1a02b6462f0441ffbadbb8188b1bd9c0f8174fecd9308f",
+      "hf_dtype": "bfloat16",
+      "hf_sha256": "0960438e0541dfa529703f5e2b37565aec66b354f6e45bfdf121ecadfcd86dd4",
+      "parameter_count": 508800,
+      "tensor_type_histogram": {
+        "F32": 9,
+        "Q4_0": 15
+      },
+      "validation": "passed"
+    },
+    {
+      "architecture": "smollm3",
+      "case": "smollm3-random-model",
+      "eog_token_ids": [
+        1
+      ],
+      "gguf_quantization": "Q4_0",
+      "gguf_sha256": "17a686b0b8c6ac17fd051bbe361570b90bed5127948d55def3df66db113b500c",
+      "hf_dtype": "bfloat16",
+      "hf_sha256": "f4cd1bbca5840cfc4d769e04a4ee1fd2c8609b103449a4a351451a6ddf4a7f07",
+      "parameter_count": 4917504,
+      "tensor_type_histogram": {
+        "F32": 9,
+        "Q4_0": 29
+      },
+      "validation": "passed"
+    }
+  ],
+  "llama_cpp_commit": "40b740ad05c531b9d57aca6698c3ed553a9e784c",
+  "schema_version": 1
+}

README.md

@@ -7,6 +7,7 @@ tags:
 - llama
 - gemma4
 - qwen3
+- smollm3
 - transformers
 - safetensors
 - gguf
@@ -17,7 +18,7 @@ tags:
 
 # Deterministic Random Models
 
-This dataset contains nine small, deterministic language-model fixtures for
+This dataset contains ten small, deterministic language-model fixtures for
 model-format, loader, inference, compatibility, and conformance testing. They
 are not trained models and must not be used for language-model quality
 evaluation.
@@ -38,10 +39,11 @@ checkpoint weights are included.
 | `livekit-turn-detector` | Llama | 132,336 | F32 | Q4_K_M | explicit head dimension, GQA |
 | `gemma4-random-model` | Gemma 4 | 1,519,168 | BF16 | Q4_0 | five-local/one-global attention schedule |
 | `qwen3-random-model` | Qwen 3 | 508,800 | BF16 | Q4_0 | wide Q projection and Q/K head norms |
+| `smollm3-random-model` | SmolLM3 | 4,917,504 | BF16 | Q4_0 | three-RoPE/one-no-RoPE layer schedule |
 
 The seven Llama cases are derived from real Hugging Face configuration files by
-a preservation-first shrinker. Gemma 4 and Qwen 3 cannot be reduced by selecting
-each field independently, so their tiny geometries preserve architecture-specific
+a preservation-first shrinker. Gemma 4, Qwen 3, and SmolLM3 retain
+architecture-specific reduced geometries that preserve important
 ratios, tensor inventories, and layer schedules observed in locally downloaded
 upstream GGUF models. Published case names use `random-model` rather than
 `tiny-model` to avoid collision with a separately maintained TinyModel collection.
@@ -85,7 +87,7 @@ The architecture-specific cases use:
 `-- metadata.json
 ```
 
-`manifest.json` is the machine-readable index of all nine model packages and
+`manifest.json` is the machine-readable index of all ten model packages and
 their SHA-256 hashes and sizes.
 
 ## Synthetic weights and tokenizers
@@ -116,7 +118,7 @@ F32 or BF16 reference exactly.
 
 ## Reproducibility and scope
 
-The Hugging Face weights, configs, and GGUF outputs for Gemma 4 and Qwen 3 were
+The Hugging Face weights, configs, and GGUF outputs for Gemma 4, Qwen 3, and SmolLM3 were
 independently regenerated and found byte-identical. The Llama cases retain their
 source revisions, source-config hashes, shrink decisions, and generation
 provenance in each case directory.
@@ -127,4 +129,4 @@ and metadata files may remain subject to terms of their respective upstream
 repositories; consult their recorded provenance before redistribution.
 
 See `REPORT.md`, `GGUF_Q4_K_M_REPORT.json`, and
-`GEMMA4_QWEN3_REPORT.json` for collection-level summaries.
+`ARCHITECTURE_RANDOM_MODELS_REPORT.json` for collection-level summaries.

SHA256SUMS

@@ -1,7 +1,7 @@
-a9498453b9883cadce5862a6d8a862949d8a49b3690a5a88673cea1b7b866b42  GEMMA4_QWEN3_REPORT.json
+c9732b21b282ff81879aa6eb2a61784a14825331847b62384da6d0a81f7218ea  ARCHITECTURE_RANDOM_MODELS_REPORT.json
 18ec0cdad98853d7f10381790713f0b3190bd4f24962594bf576e11fdb07f76f  GGUF_Q4_K_M_ADDED_FOUR_REPORT.json
 695036ee0e223d3d79170fb3a0f795eb88f6337fb2e29c18f0b549979c057208  GGUF_Q4_K_M_REPORT.json
-5e8d5eeee84a35177cba72fe8487d04295d2a9c52eb790b16ef5414eabbbcf65  README.md
+9c7a6588b7355830d75c57171a9368a7db87d192b0f05443db1233199cd893b6  README.md
 95d67a407dc96839250fe23e85e0a5bab9d355ee463b0c87bf9e33164c952b52  REPORT.md
 02130c5ef4fe4e00bf26cc6e0eb288b90916acf827031bf572ffc02a48d5f3f0  deepseek-coder/case.json
 616662945b2382e3c4ef36c05b40ebacea17e2e16fa3a723b7fd49eb5d0e9fdd  deepseek-coder/config-diff.json
@@ -72,7 +72,7 @@ f003d32e36ddc9d7c7a50c186da5c544b2828a0d04ca62aeac741d765727a635  livekit-turn-d
 564a5f0d3380d83d0ffcfa509a5e5369c0610f586cb6b3d0d45ccaee76610327  livekit-turn-detector/tokenizer/tokenizer_config.json
 7b1258e1304c5f09f3bc64cd4b569e7ddfa55a062c0581d0e5cb1a3d7a17fcc8  livekit-turn-detector/tokenizer/vocabulary.json
 391312ee129392e2b35b6a238593ca5ddff3dd51c6c324135f1bb0114be1e068  livekit-turn-detector/validation.json
-4ef64a4405f5f193022a8ecbc70513ac49278ac6b1ce2f101a98c8bf56f4c9a2  manifest.json
+853c059a6216e5edc98ad1f444fb4dd010d263ebef36a7e5ca3a3ff715c3ab36  manifest.json
 047c4545232dc055c2c7e4a30a70c22e0e75e9e777e539732128f65c32b0c1cb  minicpm5/case.json
 282a55aae55b59573109fe37222197f298600ad691573bcd84ea5d4dc1afcc0a  minicpm5/config-diff.json
 d088266a01cd2f9a6107877932d6af9f65e6d00202292226eab76af8ea0e889f  minicpm5/gguf/metadata.json
@@ -143,6 +143,20 @@ d4a932badf1309a3bf52a5711d6b614276f54c7985fcf3510ae6c67f75a16967  smollm2-instru
 2932a00a19894e082de22185578d4535c33f8893f4987cdeb913ec95d08f3401  smollm2-instruct/tokenizer/tokenizer_config.json
 cf5fd07e0df593dc5a43ae647f1f3d1898588a5ec6b5af6ea967cf82be72fd99  smollm2-instruct/tokenizer/vocabulary.json
 62f70e96ae4a7b114b522c412c131b2710f85bae174d2630bc4d8879028a7483  smollm2-instruct/validation.json
+306151174d9d84dc2b3e977f774b3a50571c14fc81c6239cd1b84e18b1c38089  smollm3-random-model/CONFIG_DECISION.md
+77a1f80f5a19edbfcb9c6f778188d9a3e0beddcc3fa0089b7b85805a9caa2dc1  smollm3-random-model/gguf-q4_0/convert.log
+eb0c84b3bbaac985ddfb8a1831f1070817065a29f1ac4bf290112c26e48e68b1  smollm3-random-model/gguf-q4_0/quantize.log
+17a686b0b8c6ac17fd051bbe361570b90bed5127948d55def3df66db113b500c  smollm3-random-model/gguf-q4_0/smollm3-random-model-Q4_0.gguf
+f9a288571e5de8ddbc176aaa35ef7422e0d6b213534e70103138f9573faba8d5  smollm3-random-model/hf-bf16/config.json
+7d70f6a71b49e15e5116e1a76953b9cbb6c611f5e811bc6e3f96056607fa341f  smollm3-random-model/hf-bf16/generation_config.json
+f4cd1bbca5840cfc4d769e04a4ee1fd2c8609b103449a4a351451a6ddf4a7f07  smollm3-random-model/hf-bf16/model.safetensors
+ba2df415c6a5595f4e9227a29d8e8b30d4d3ff6b9af7243005f191373868af45  smollm3-random-model/hf-bf16/tokenizer.json
+0f88117a2f8c52a9eabe2d61f7301b6f9b58b7a56644888da4ec9970b1efbcee  smollm3-random-model/hf-bf16/tokenizer.model
+d0fab7832fecaac8554eaebfad305d55655cc79b3404eb9a7b0a4fc9dce08a01  smollm3-random-model/hf-bf16/tokenizer_config.json
+9ef6aa7b13679b03a6010fa101d55e082a656469b48f29f78990dcf8c21aa1db  smollm3-random-model/metadata.json
+4444092a55ff192344629868922c12ed25ed6fd070b4e8f5326a361a2728b3be  smollm3-random-model/reference/gguf-native.json
+ddcd5c412fda5119791b4ef082edbc56f8b207b312fca28c92929f5e1e74e314  smollm3-random-model/reference/hf-outputs.safetensors
+719a2f5ea966e92e42ebcc43dd73f0714b5d740c9f426d22c942dbb1e7efb6ab  smollm3-random-model/reference/inputs.json
 8ffefc53e4e83fb5e5cba25e59f8f8246ac38243769e8d3b6266144ef6aa9529  tinyllama-chat/case.json
 30b65ef261fed6560ea3c0073aade016c4285ef1c4a372921d01e4b29bcb2ecf  tinyllama-chat/config-diff.json
 9ec815048e6120c0d82b1e073891801c57ab89ba04dadc835a99b31282169284  tinyllama-chat/gguf/metadata.json

manifest.json

@@ -172,6 +172,26 @@
           "size_bytes": 1020120
         }
       }
+    },
+    {
+      "architecture": "smollm3",
+      "auxiliary_tokenizer": {
+        "path": "smollm3-random-model/hf-bf16/tokenizer.json",
+        "sha256": "ba2df415c6a5595f4e9227a29d8e8b30d4d3ff6b9af7243005f191373868af45"
+      },
+      "case": "smollm3-random-model",
+      "packages": {
+        "gguf_q4_0": {
+          "path": "smollm3-random-model/gguf-q4_0/smollm3-random-model-Q4_0.gguf",
+          "sha256": "17a686b0b8c6ac17fd051bbe361570b90bed5127948d55def3df66db113b500c",
+          "size_bytes": 2780064
+        },
+        "safetensors_bf16": {
+          "path": "smollm3-random-model/hf-bf16/model.safetensors",
+          "sha256": "f4cd1bbca5840cfc4d769e04a4ee1fd2c8609b103449a4a351451a6ddf4a7f07",
+          "size_bytes": 9839096
+        }
+      }
     }
   ],
   "schema_version": 1

smollm3-random-model/CONFIG_DECISION.md

@@ -0,0 +1,29 @@
+# SmolLM3 random-model configuration decision
+
+The source package at `/home/codex/llm_models/hfmodels/SmolLM3-3B` identifies
+itself as `SmolLM3ForCausalLM` with 3,075,098,624 parameters. The generated
+fixture is published as `smollm3-random-model`.
+
+The source has hidden width 2048, FFN width 11008, 36 layers, 16 query heads,
+four KV heads, context 65536, RoPE base 5000000, and tied embeddings. Its
+attention schedule repeats three RoPE layers followed by one no-RoPE layer.
+
+The reduced geometry uses hidden width 256, FFN width 1376, four layers, four
+query heads, one KV head, head dimension 64, context 128, and vocabulary 128.
+This preserves:
+
+- the query/KV head ratio of four;
+- the exact FFN ratio of 43/8;
+- one complete three-RoPE/one-no-RoPE schedule;
+- RoPE base, RMS epsilon, bias settings, and tied embeddings;
+- matrix quantization axes divisible by 32 for Q4_0.
+
+The official Transformers 5.14.1 `SmolLM3ForCausalLM` construction contains
+4,917,504 trainable parameters. The retained Hugging Face package is BF16. The
+GGUF path uses the pinned official llama.cpp SmolLM3 converter, a temporary F16
+GGUF, and Q4_0 quantization.
+
+The Hugging Face package contains the deterministic 128-token `tokenizer.json`.
+For official converter compatibility it also carries the existing 128-token
+SentencePiece auxiliary artifact whose EOG set is `{1}`. Both are synthetic;
+direct token IDs are the numerical reference interface.

smollm3-random-model/gguf-q4_0/convert.log

@@ -0,0 +1,68 @@
+INFO:hf-to-gguf:Loading model: hf-bf16
+INFO:hf-to-gguf:Model architecture: SmolLM3ForCausalLM
+INFO:hf-to-gguf:gguf: indexing model part 'model.safetensors'
+INFO:gguf.gguf_writer:gguf: This GGUF file is for Little Endian only
+INFO:hf-to-gguf:Exporting model...
+INFO:hf-to-gguf:token_embd.weight,          torch.bfloat16 --> F16, shape = {256, 128}
+INFO:hf-to-gguf:blk.0.attn_norm.weight,     torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.0.ffn_down.weight,      torch.bfloat16 --> F16, shape = {1376, 256}
+INFO:hf-to-gguf:blk.0.ffn_gate.weight,      torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.0.ffn_up.weight,        torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.0.ffn_norm.weight,      torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.0.attn_k.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.0.attn_output.weight,   torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.0.attn_q.weight,        torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.0.attn_v.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.1.attn_norm.weight,     torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.1.ffn_down.weight,      torch.bfloat16 --> F16, shape = {1376, 256}
+INFO:hf-to-gguf:blk.1.ffn_gate.weight,      torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.1.ffn_up.weight,        torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.1.ffn_norm.weight,      torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.1.attn_k.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.1.attn_output.weight,   torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.1.attn_q.weight,        torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.1.attn_v.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.2.attn_norm.weight,     torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.2.ffn_down.weight,      torch.bfloat16 --> F16, shape = {1376, 256}
+INFO:hf-to-gguf:blk.2.ffn_gate.weight,      torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.2.ffn_up.weight,        torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.2.ffn_norm.weight,      torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.2.attn_k.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.2.attn_output.weight,   torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.2.attn_q.weight,        torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.2.attn_v.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.3.attn_norm.weight,     torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.3.ffn_down.weight,      torch.bfloat16 --> F16, shape = {1376, 256}
+INFO:hf-to-gguf:blk.3.ffn_gate.weight,      torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.3.ffn_up.weight,        torch.bfloat16 --> F16, shape = {256, 1376}
+INFO:hf-to-gguf:blk.3.ffn_norm.weight,      torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:blk.3.attn_k.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:blk.3.attn_output.weight,   torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.3.attn_q.weight,        torch.bfloat16 --> F16, shape = {256, 256}
+INFO:hf-to-gguf:blk.3.attn_v.weight,        torch.bfloat16 --> F16, shape = {256, 64}
+INFO:hf-to-gguf:output_norm.weight,         torch.bfloat16 --> F32, shape = {256}
+INFO:hf-to-gguf:Set meta model
+INFO:hf-to-gguf:Set model parameters
+INFO:hf-to-gguf:gguf: context length = 128
+INFO:hf-to-gguf:gguf: embedding length = 256
+INFO:hf-to-gguf:gguf: feed forward length = 1376
+INFO:hf-to-gguf:gguf: head count = 4
+INFO:hf-to-gguf:gguf: key-value head count = 1
+WARNING:hf-to-gguf:Unknown RoPE type: default
+INFO:hf-to-gguf:gguf: rope scaling type = NONE
+INFO:hf-to-gguf:gguf: rope theta = 5000000.0
+INFO:hf-to-gguf:gguf: rms norm epsilon = 1e-06
+INFO:hf-to-gguf:gguf: file type = 1
+INFO:hf-to-gguf:Set model quantization version
+INFO:hf-to-gguf:Set model tokenizer
+INFO:gguf.vocab:Setting special token type bos to 2
+INFO:gguf.vocab:Setting special token type eos to 1
+INFO:gguf.vocab:Setting special token type unk to 3
+INFO:gguf.vocab:Setting special token type pad to 0
+INFO:gguf.vocab:Setting special token type mask to 4
+INFO:gguf.gguf_writer:Writing the following files:
+INFO:gguf.gguf_writer:/home/codex/tmp/smollm3-random-work-20260810/smollm3-random-model-F16.gguf: n_tensors = 38, total_size = 9.8M
+
+Writing:   0%|          | 0.00/9.84M [00:00<?, ?byte/s]
+Writing: 100%|██████████| 9.84M/9.84M [00:00<00:00, 373Mbyte/s]
+INFO:hf-to-gguf:Model successfully exported to /home/codex/tmp/smollm3-random-work-20260810/smollm3-random-model-F16.gguf

smollm3-random-model/gguf-q4_0/quantize.log

@@ -0,0 +1,76 @@
+llama_print_build_info: build = 1 (40b740a)
+llama_print_build_info: built with Clang 21.1.8 for Linux x86_64
+llama_quantize: quantizing '/home/codex/tmp/smollm3-random-work-20260810/smollm3-random-model-F16.gguf' to '/home/codex/conf_track/artifacts/smollm3-v0/smollm3-random-model/gguf-q4_0/smollm3-random-model-Q4_0.gguf' as Q4_0
+llama_model_loader: loaded meta data with 26 key-value pairs and 38 tensors from /home/codex/tmp/smollm3-random-work-20260810/smollm3-random-model-F16.gguf (version GGUF V3 (latest))
+llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
+llama_model_loader: - kv   0:                       general.architecture str              = smollm3
+llama_model_loader: - kv   1:                               general.type str              = model
+llama_model_loader: - kv   2:                               general.name str              = smollm3-random-model
+llama_model_loader: - kv   3:                         general.size_label str              = 4.9M
+llama_model_loader: - kv   4:                        smollm3.block_count u32              = 4
+llama_model_loader: - kv   5:                     smollm3.context_length u32              = 128
+llama_model_loader: - kv   6:                   smollm3.embedding_length u32              = 256
+llama_model_loader: - kv   7:                smollm3.feed_forward_length u32              = 1376
+llama_model_loader: - kv   8:               smollm3.attention.head_count u32              = 4
+llama_model_loader: - kv   9:            smollm3.attention.head_count_kv u32              = 1
+llama_model_loader: - kv  10:                     smollm3.rope.freq_base f32              = 5000000.000000
+llama_model_loader: - kv  11:   smollm3.attention.layer_norm_rms_epsilon f32              = 0.000001
+llama_model_loader: - kv  12:                          general.file_type u32              = 1
+llama_model_loader: - kv  13:                         smollm3.vocab_size u32              = 128
+llama_model_loader: - kv  14:               smollm3.rope.dimension_count u32              = 64
+llama_model_loader: - kv  15:               general.quantization_version u32              = 2
+llama_model_loader: - kv  16:                       tokenizer.ggml.model str              = llama
+llama_model_loader: - kv  17:                         tokenizer.ggml.pre str              = default
+llama_model_loader: - kv  18:                      tokenizer.ggml.tokens arr[str,128]     = ["<unk>", "<s>", "<reserved02>", "<au...
+llama_model_loader: - kv  19:                      tokenizer.ggml.scores arr[f32,128]     = [0.000000, 0.000000, 0.000000, 0.0000...
+llama_model_loader: - kv  20:                  tokenizer.ggml.token_type arr[i32,128]     = [2, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
+llama_model_loader: - kv  21:                tokenizer.ggml.bos_token_id u32              = 2
+llama_model_loader: - kv  22:                tokenizer.ggml.eos_token_id u32              = 1
+llama_model_loader: - kv  23:            tokenizer.ggml.unknown_token_id u32              = 3
+llama_model_loader: - kv  24:            tokenizer.ggml.padding_token_id u32              = 0
+llama_model_loader: - kv  25:               tokenizer.ggml.mask_token_id u32              = 4
+llama_model_loader: - type  f32:    9 tensors
+llama_model_loader: - type  f16:   29 tensors
+[   1/  38] output_norm.weight                   - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[   2/  38] token_embd.weight                    - [   256,    128,      1,      1], type =    f16, converting to q4_0 .. size =     0.06 MiB ->     0.02 MiB
+[   3/  38] blk.0.attn_k.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[   4/  38] blk.0.attn_norm.weight               - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[   5/  38] blk.0.attn_output.weight             - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[   6/  38] blk.0.attn_q.weight                  - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[   7/  38] blk.0.attn_v.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[   8/  38] blk.0.ffn_down.weight                - [  1376,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[   9/  38] blk.0.ffn_gate.weight                - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  10/  38] blk.0.ffn_norm.weight                - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  11/  38] blk.0.ffn_up.weight                  - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  12/  38] blk.1.attn_k.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  13/  38] blk.1.attn_norm.weight               - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  14/  38] blk.1.attn_output.weight             - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  15/  38] blk.1.attn_q.weight                  - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  16/  38] blk.1.attn_v.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  17/  38] blk.1.ffn_down.weight                - [  1376,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  18/  38] blk.1.ffn_gate.weight                - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  19/  38] blk.1.ffn_norm.weight                - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  20/  38] blk.1.ffn_up.weight                  - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  21/  38] blk.2.attn_k.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  22/  38] blk.2.attn_norm.weight               - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  23/  38] blk.2.attn_output.weight             - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  24/  38] blk.2.attn_q.weight                  - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  25/  38] blk.2.attn_v.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  26/  38] blk.2.ffn_down.weight                - [  1376,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  27/  38] blk.2.ffn_gate.weight                - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  28/  38] blk.2.ffn_norm.weight                - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  29/  38] blk.2.ffn_up.weight                  - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  30/  38] blk.3.attn_k.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  31/  38] blk.3.attn_norm.weight               - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  32/  38] blk.3.attn_output.weight             - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  33/  38] blk.3.attn_q.weight                  - [   256,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.12 MiB ->     0.04 MiB
+[  34/  38] blk.3.attn_v.weight                  - [   256,     64,      1,      1], type =    f16, converting to q4_0 .. size =     0.03 MiB ->     0.01 MiB
+[  35/  38] blk.3.ffn_down.weight                - [  1376,    256,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  36/  38] blk.3.ffn_gate.weight                - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+[  37/  38] blk.3.ffn_norm.weight                - [   256,      1,      1,      1], type =    f32, size =    0.001 MiB
+[  38/  38] blk.3.ffn_up.weight                  - [   256,   1376,      1,      1], type =    f16, converting to q4_0 .. size =     0.67 MiB ->     0.19 MiB
+llama_model_quantize_impl: model size  =     9.38 MiB (16.01 BPW)
+llama_model_quantize_impl: quant size  =     2.65 MiB (4.51 BPW)
+
+llama_quantize: quantize time =    42.99 ms
+llama_quantize:    total time =    42.99 ms

smollm3-random-model/gguf-q4_0/smollm3-random-model-Q4_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17a686b0b8c6ac17fd051bbe361570b90bed5127948d55def3df66db113b500c
+size 2780064

smollm3-random-model/hf-bf16/config.json

@@ -0,0 +1,47 @@
+{
+  "architectures": [
+    "SmolLM3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "dtype": "bfloat16",
+  "eos_token_id": 1,
+  "hidden_act": "silu",
+  "hidden_size": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 1376,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 128,
+  "max_window_layers": 28,
+  "mlp_bias": false,
+  "model_type": "smollm3",
+  "no_rope_layer_interval": 4,
+  "no_rope_layers": [
+    1,
+    1,
+    1,
+    0
+  ],
+  "num_attention_heads": 4,
+  "num_hidden_layers": 4,
+  "num_key_value_heads": 1,
+  "pad_token_id": 0,
+  "pretraining_tp": 2,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 5000000.0,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.14.1",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 128
+}

smollm3-random-model/hf-bf16/generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "output_attentions": false,
+  "output_hidden_states": false,
+  "pad_token_id": 0,
+  "transformers_version": "5.14.1",
+  "use_cache": false
+}

smollm3-random-model/hf-bf16/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4cd1bbca5840cfc4d769e04a4ee1fd2c8609b103449a4a351451a6ddf4a7f07
+size 9839096

smollm3-random-model/hf-bf16/tokenizer.json

@@ -0,0 +1,237 @@
+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "<pad>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 1,
+      "content": "<eos>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 2,
+      "content": "<bos>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 3,
+      "content": "<unk>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 4,
+      "content": "<mask>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": null,
+  "pre_tokenizer": {
+    "type": "ByteLevel",
+    "add_prefix_space": false,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "post_processor": {
+    "type": "TemplateProcessing",
+    "single": [
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      }
+    ],
+    "pair": [
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "B",
+          "type_id": 1
+        }
+      }
+    ],
+    "special_tokens": {}
+  },
+  "decoder": {
+    "type": "ByteLevel",
+    "add_prefix_space": true,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "model": {
+    "type": "BPE",
+    "dropout": null,
+    "unk_token": "<unk>",
+    "continuing_subword_prefix": null,
+    "end_of_word_suffix": null,
+    "fuse_unk": false,
+    "byte_fallback": false,
+    "ignore_merges": false,
+    "vocab": {
+      "<pad>": 0,
+      "<eos>": 1,
+      "<bos>": 2,
+      "<unk>": 3,
+      "<mask>": 4,
+      "a": 5,
+      "b": 6,
+      "ab": 7,
+      "<t008>": 8,
+      "<t009>": 9,
+      "<t010>": 10,
+      "<t011>": 11,
+      "<t012>": 12,
+      "<t013>": 13,
+      "<t014>": 14,
+      "<t015>": 15,
+      "<t016>": 16,
+      "<t017>": 17,
+      "<t018>": 18,
+      "<t019>": 19,
+      "<t020>": 20,
+      "<t021>": 21,
+      "<t022>": 22,
+      "<t023>": 23,
+      "<t024>": 24,
+      "<t025>": 25,
+      "<t026>": 26,
+      "<t027>": 27,
+      "<t028>": 28,
+      "<t029>": 29,
+      "<t030>": 30,
+      "<t031>": 31,
+      "<t032>": 32,
+      "<t033>": 33,
+      "<t034>": 34,
+      "<t035>": 35,
+      "<t036>": 36,
+      "<t037>": 37,
+      "<t038>": 38,
+      "<t039>": 39,
+      "<t040>": 40,
+      "<t041>": 41,
+      "<t042>": 42,
+      "<t043>": 43,
+      "<t044>": 44,
+      "<t045>": 45,
+      "<t046>": 46,
+      "<t047>": 47,
+      "<t048>": 48,
+      "<t049>": 49,
+      "<t050>": 50,
+      "<t051>": 51,
+      "<t052>": 52,
+      "<t053>": 53,
+      "<t054>": 54,
+      "<t055>": 55,
+      "<t056>": 56,
+      "<t057>": 57,
+      "<t058>": 58,
+      "<t059>": 59,
+      "<t060>": 60,
+      "<t061>": 61,
+      "<t062>": 62,
+      "<t063>": 63,
+      "<t064>": 64,
+      "<t065>": 65,
+      "<t066>": 66,
+      "<t067>": 67,
+      "<t068>": 68,
+      "<t069>": 69,
+      "<t070>": 70,
+      "<t071>": 71,
+      "<t072>": 72,
+      "<t073>": 73,
+      "<t074>": 74,
+      "<t075>": 75,
+      "<t076>": 76,
+      "<t077>": 77,
+      "<t078>": 78,
+      "<t079>": 79,
+      "<t080>": 80,
+      "<t081>": 81,
+      "<t082>": 82,
+      "<t083>": 83,
+      "<t084>": 84,
+      "<t085>": 85,
+      "<t086>": 86,
+      "<t087>": 87,
+      "<t088>": 88,
+      "<t089>": 89,
+      "<t090>": 90,
+      "<t091>": 91,
+      "<t092>": 92,
+      "<t093>": 93,
+      "<t094>": 94,
+      "<t095>": 95,
+      "<t096>": 96,
+      "<t097>": 97,
+      "<t098>": 98,
+      "<t099>": 99,
+      "<t100>": 100,
+      "<t101>": 101,
+      "<t102>": 102,
+      "<t103>": 103,
+      "<t104>": 104,
+      "<t105>": 105,
+      "<t106>": 106,
+      "<t107>": 107,
+      "<t108>": 108,
+      "<t109>": 109,
+      "<t110>": 110,
+      "<t111>": 111,
+      "<t112>": 112,
+      "<t113>": 113,
+      "<t114>": 114,
+      "<t115>": 115,
+      "<t116>": 116,
+      "<t117>": 117,
+      "<t118>": 118,
+      "<t119>": 119,
+      "<t120>": 120,
+      "<t121>": 121,
+      "<t122>": 122,
+      "<t123>": 123,
+      "<t124>": 124,
+      "<t125>": 125,
+      "<t126>": 126,
+      "<t127>": 127
+    },
+    "merges": [
+      [
+        "a",
+        "b"
+      ]
+    ]
+  }
+}

smollm3-random-model/hf-bf16/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f88117a2f8c52a9eabe2d61f7301b6f9b58b7a56644888da4ec9970b1efbcee
+size 1980

smollm3-random-model/hf-bf16/tokenizer_config.json

@@ -0,0 +1,10 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<bos>",
+  "eos_token": "<eos>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": "<unk>"
+}

smollm3-random-model/metadata.json

@@ -0,0 +1,91 @@
+{
+  "architecture": "smollm3",
+  "gguf": {
+    "comparison_to_hf_bf16": {
+      "decode_logits": {
+        "max_absolute_error": 0.09191668033599854,
+        "mean_absolute_error": 0.023608334362506866,
+        "top1_agreement": true,
+        "top5_overlap": 4
+      },
+      "policy": "informational; quantized and BF16 executions are not required to be byte-identical",
+      "prefill_last_logits": {
+        "max_absolute_error": 0.09857787936925888,
+        "mean_absolute_error": 0.023846685886383057,
+        "top1_agreement": true,
+        "top5_overlap": 4
+      }
+    },
+    "converter_command": [
+      "PYTHON",
+      "CONVERTER",
+      "HF_PACKAGE",
+      "--outfile",
+      "F16_OUTPUT",
+      "--outtype",
+      "f16",
+      "--model-name",
+      "smollm3-random-model"
+    ],
+    "direct_token_validation": {
+      "all_logits_finite": true,
+      "decode_token_id": 17,
+      "eog_token_ids": [
+        1
+      ],
+      "prefill_token_ids": [
+        5,
+        7,
+        11,
+        13
+      ],
+      "repeated_output_byte_identical": true,
+      "status": "passed"
+    },
+    "f16_intermediate_retained_in_package": false,
+    "f16_intermediate_sha256": "f786a25e21c99c4d3fb33cf672fcab42bdc3ac132f2071de717073636c4c2680",
+    "model_sha256": "17a686b0b8c6ac17fd051bbe361570b90bed5127948d55def3df66db113b500c",
+    "model_size_bytes": 2780064,
+    "quantization": "Q4_0",
+    "quantizer_command": [
+      "LLAMA_QUANTIZE",
+      "--token-embedding-type",
+      "q4_0",
+      "F16_INPUT",
+      "OUTPUT",
+      "Q4_0"
+    ],
+    "tensor_type_histogram": {
+      "F32": 9,
+      "Q4_0": 29
+    }
+  },
+  "hf": {
+    "config_sha256": "f9a288571e5de8ddbc176aaa35ef7422e0d6b213534e70103138f9573faba8d5",
+    "construct_prefill_decode_reload": "passed",
+    "converter_sentencepiece_sha256": "0f88117a2f8c52a9eabe2d61f7301b6f9b58b7a56644888da4ec9970b1efbcee",
+    "dtype": "bfloat16",
+    "model_sha256": "f4cd1bbca5840cfc4d769e04a4ee1fd2c8609b103449a4a351451a6ddf4a7f07",
+    "reload_tolerance": {
+      "atol": 0.02,
+      "rtol": 0.02
+    },
+    "tokenizer_json_sha256": "ba2df415c6a5595f4e9227a29d8e8b30d4d3ff6b9af7243005f191373868af45",
+    "transformers_version": "5.14.1"
+  },
+  "parameter_count": 4917504,
+  "reproducibility": {
+    "gguf_q4_0_byte_identical": true,
+    "hf_config_byte_identical": true,
+    "hf_model_byte_identical": true,
+    "native_repeated_output_byte_identical": true
+  },
+  "rng": {
+    "increment": 1442695040888963407,
+    "multiplier": 6364136223846793005,
+    "name": "tlfloat LCG64 equation",
+    "warmup_steps": 10
+  },
+  "schema_version": 1,
+  "seed": 6002541067470975745
+}

smollm3-random-model/reference/gguf-native.json

@@ -0,0 +1,10 @@
+{
+  "schema_version": 1,
+  "vocab_size": 128,
+  "prefill_token_ids": [5,7,11,13],
+  "decode_token_id": 17,
+  "eog_token_ids": [1],
+  "all_logits_finite": true,
+  "prefill_last_logits": [0.350657523,0.245601147,0.185828045,0.0872669816,-0.248535678,0.216705814,0.098706007,0.067492865,0.0902635753,0.0860311389,0.287080705,-0.192821488,0.0804038569,0.636609077,0.240223914,0.0575771704,-0.170419544,0.104185969,-0.120821357,-0.0154661946,0.31676352,-0.0446437784,-0.0388852209,-0.0993875638,-0.341283709,0.0729972646,0.433401316,-0.27149269,0.254691511,-0.0325341113,-0.253310263,0.139225543,-0.00844951533,-0.144148409,0.381891876,0.0964927375,0.00417396333,0.0205155089,-0.468620837,-0.0534039363,-0.0309029669,-0.0159435179,-0.191765785,-0.0309073832,0.060901247,-0.0437048823,-0.272850752,0.0477128029,-0.147300392,-0.0993154123,-0.30048281,0.154198244,-0.0939248279,0.023117898,0.191344813,0.0888454318,0.0576772429,0.0138832312,-0.173797771,-0.178063288,0.0579373911,0.0921991915,0.0546639301,-0.0848264918,0.273745567,0.0871551633,0.16084519,-0.237837672,-0.188799113,0.0468373187,-0.136669695,0.339535147,-0.0617689639,0.186668202,-0.0222822782,0.0100252936,0.00999814644,0.0238801166,0.174783677,0.0882279649,0.134765789,-0.00743198581,-0.0283377115,-0.174449503,-0.177547276,0.0928283706,0.0965974331,-0.345378071,0.108685277,0.0226041302,-0.186054453,-0.0744341537,-0.0869331285,-0.0980709717,0.0390793867,0.298965454,0.127674684,-0.0902414173,0.184343874,-0.173850864,-0.00467307493,0.114426896,0.178143948,0.145223439,-0.22824049,-0.0174845606,0.0152879208,-0.187148482,0.0576404557,0.195420101,-0.0546050034,0.404827625,-0.60698694,0.28225562,-0.107594267,-0.0409816131,-0.221593961,-0.153167903,0.0506973863,0.20176214,0.0392604545,-0.447950959,0.0143984454,0.23511833,-0.377050936,0.438002795,0.0612642206,0.165355772],
+  "decode_logits": [0.361390561,0.259690762,0.193840697,0.0224308651,-0.180631235,0.181348011,0.0344774835,0.0258460268,0.0892843306,0.08898516,0.248940408,-0.144477874,0.100970209,0.0473254919,0.101009533,0.1632687,-0.189144939,0.689852178,-0.0114180446,-0.0392280668,0.430228204,-0.26826939,-0.0187434368,-0.0778454617,-0.31398192,0.0945572257,0.455448389,-0.353762329,0.246169433,0.0822324529,-0.0832768232,0.201291218,-0.0322698653,-0.00652883667,0.248189881,0.0995574594,0.0214569774,0.0370474011,-0.2733666,0.00299273594,-0.101147383,-0.107943103,-0.14415054,0.084869355,0.0267817583,-0.120424025,-0.214804426,0.0638376325,-0.123272434,-0.120914482,-0.281999916,0.0949365199,-0.0883979648,-0.0321664251,0.19711931,0.130908251,0.070493713,-0.0173004754,-0.211316571,-0.230640933,-0.0688304082,0.0668021739,0.0223920215,-0.0961654112,0.175528616,-0.029617073,0.225180477,-0.0793279633,-0.169067502,-0.0724712908,-0.24525097,0.282416493,-0.102731496,0.131909415,0.111945592,-0.0441522077,-0.044923,0.0778397098,0.229346365,0.135500953,0.15436694,-0.0073002018,-0.0653385445,-0.173108846,-0.202871338,0.024294259,0.010343954,-0.339766681,0.0781216472,0.0684290603,-0.159132197,-0.0493372455,0.0273793489,-0.123151906,0.0259634107,0.193071097,0.147674754,-0.0749316886,0.274478108,-0.153305024,-0.0489038937,0.159410432,0.168823883,0.104770966,-0.168470278,0.12241514,0.0164226107,-0.299293399,0.130618289,0.188386604,0.00230943272,0.41047129,-0.566889167,0.322926939,-0.0434842482,-0.116828635,-0.273065031,-0.0706957728,-0.0685233772,0.152179644,0.0153105902,-0.399265587,0.12370012,0.0944165736,-0.206045151,0.445940405,0.00132466271,0.0807432383]
+}

smollm3-random-model/reference/hf-outputs.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddcd5c412fda5119791b4ef082edbc56f8b207b312fca28c92929f5e1e74e314
+size 2720

smollm3-random-model/reference/inputs.json

@@ -0,0 +1,9 @@
+{
+  "decode_token_id": 17,
+  "prefill_token_ids": [
+    5,
+    7,
+    11,
+    13
+  ]
+}