Qwen3.8-27B-FP8 / validation /validation_report.json
Chungulus's picture
Add FP8 validation evidence
6550026 verified
Raw
History Blame Contribute Delete
10.5 kB
{
"target": "FP8",
"artifact": "/mnt/c/Users/chunk/OneDrive/Documents/quant3.8/qwen38-cuda-quants/outputs/Qwen3.8-27B-FP8",
"started_utc": "2026-08-18T06:44:09+00:00",
"smoke_test": false,
"small_architecture_test": false,
"resume_evidence": {
"path": "/mnt/c/Users/chunk/OneDrive/Documents/quant3.8/qwen38-cuda-quants/logs/full_fp8_validate_resume3.console.log",
"sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c",
"passing_generation_cases": [
"coding",
"debugging",
"eos",
"factual",
"json",
"long_context",
"math",
"multilingual",
"multiturn",
"thinking_off"
],
"passing_tool_cases": [
"chained",
"missing_required",
"nested",
"weather"
]
},
"checks": {
"config": {
"passed": true,
"class": "Qwen3_5Config",
"model_type": "qwen3_5",
"language_layers": 64,
"advertised_context_length": 262144,
"quantization_method": "fp8",
"expected_quantization_method": "fp8",
"quantization_config": {
"bits": 8,
"checkpoint_format": "float8_e4m3fn",
"desc_act": false,
"format": "float8_e4m3fn",
"group_size": -1,
"lm_head": false,
"meta": {
"act_group_aware": null,
"auto_forward_data_parallel": true,
"damp_auto_increment": null,
"damp_percent": null,
"dense_vram_strategy": "exclusive",
"dense_vram_strategy_devices": null,
"fallback": {
"smooth": null,
"strategy": "rtn",
"threshold": "0.5%"
},
"foem": null,
"gc_mode": "interval",
"gptaq": null,
"moe_vram_strategy": "exclusive",
"moe_vram_strategy_devices": null,
"mse": null,
"offload_to_disk": false,
"offload_to_disk_path": null,
"pack_impl": "cpu",
"quantizer": [
"gptqmodel:7.3.4"
],
"static_groups": null,
"true_sequential": true,
"uri": "https://github.com/modelcloud/gptqmodel",
"wait_for_submodule_finalizers": false
},
"method": "fp8",
"pack_dtype": "int32",
"quant_method": "fp8",
"weight_scale_method": "row",
"weight_scale_semantics": "inverse"
}
},
"processor": {
"passed": true,
"class": "Qwen3VLProcessor",
"chat_template_present": true,
"special_token_ids": {
"image_token_id": 248056,
"video_token_id": 248057,
"vision_start_token_id": 248053,
"vision_end_token_id": 248054
}
},
"chat_template": {
"present": true,
"sha256": "c3cf9e34abf4f9e36c2d72165aa9c132d3e2a725b6c2586aaa3a8af9d7a81041",
"matches_reference": true,
"passed": true
},
"inherited_metadata": {
"passed": true,
"files_checked": [
"generation_config.json",
"tokenizer.json",
"tokenizer_config.json",
"vocab.json",
"merges.txt",
"chat_template.jinja",
"preprocessor_config.json",
"video_preprocessor_config.json",
"processor_config.json",
"LICENSE"
],
"missing": [],
"mismatched": []
},
"tensors": {
"files": [
{
"name": "model-00001-of-00008.safetensors",
"bytes": 3516829256
},
{
"name": "model-00002-of-00008.safetensors",
"bytes": 4268842872
},
{
"name": "model-00003-of-00008.safetensors",
"bytes": 4270471840
},
{
"name": "model-00004-of-00008.safetensors",
"bytes": 4280916456
},
{
"name": "model-00005-of-00008.safetensors",
"bytes": 4265221120
},
{
"name": "model-00006-of-00008.safetensors",
"bytes": 4212755176
},
{
"name": "model-00007-of-00008.safetensors",
"bytes": 4270471904
},
{
"name": "model-00008-of-00008.safetensors",
"bytes": 2166188208
}
],
"tensor_count": 1599,
"float_tensors_scanned": 1599,
"nan_count": 0,
"inf_count": 0,
"meta_tensor_count": 0,
"has_language": true,
"has_vision": true,
"has_lm_head": true,
"has_mtp": true,
"quantized_weight_tensors": 0,
"lm_head_satisfied_by_tied_embeddings": false,
"packed_weights_required": false,
"passed": true
},
"model_load": {
"passed": true,
"loader": "GPTQModel",
"model_class": "Qwen3_5ForConditionalGeneration",
"device_map_summary": {
"cpu": 59,
"cuda:0": 9
}
},
"generation": {
"passed": true,
"cases": [
{
"name": "factual",
"prompt": "Answer in one short sentence: What is the capital of France?",
"case_passed": true,
"new_tokens": 8,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "math",
"prompt": "Compute 17 x 23. Give only the integer.",
"case_passed": true,
"new_tokens": 4,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "coding",
"prompt": "Write a Python function is_even(n) in at most three lines.",
"case_passed": true,
"new_tokens": 21,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "debugging",
"prompt": "Fix this Python expression and explain briefly: len(3).",
"case_passed": true,
"new_tokens": 96,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "multiturn",
"prompt": "The secret word is cobalt. What secret word did I just give you?",
"case_passed": true,
"new_tokens": 3,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "json",
"prompt": "Return only valid JSON with keys \"status\" and \"count\"; use status \"ok\" and count 3.",
"case_passed": true,
"new_tokens": 13,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "multilingual",
"prompt": "Responde en espanol con una frase: De que color es el cielo despejado?",
"case_passed": true,
"new_tokens": 10,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "thinking_off",
"prompt": "Give only the final answer, without reasoning: 9 + 14 = ?",
"case_passed": true,
"new_tokens": 3,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "long_context",
"prompt": "Recover the codeword from the long context and return only that word.",
"case_passed": true,
"new_tokens": 4,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"name": "eos",
"prompt": "Reply with exactly: done",
"case_passed": true,
"new_tokens": 2,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
}
]
},
"tools": {
"passed": true,
"cases": [
{
"case": "weather",
"prompt": "Use the weather tool for Chicago with unit celsius. Return only the tool call.",
"schema_passed": true,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"case": "nested",
"prompt": "Use the route tool with JSON arguments origin=A, destination=B, options={avoid_tolls:true}. Return only the tool call.",
"schema_passed": true,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"case": "chained",
"prompt": "Call lookup_user for id 7, then call list_orders using that user. Return the two tool calls.",
"schema_passed": true,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
},
{
"case": "missing_required",
"prompt": "Before calling book_trip, state which required date is missing. Do not invent it.",
"schema_passed": true,
"resumed_from_evidence": true,
"evidence_sha256": "0f648b7067b40f85ea45320b1d841c38a79bbf11f022f6e3eb4a974e57072e8c"
}
]
},
"vision": {
"text": "blue red green B 7",
"prompt_tokens": 106,
"new_tokens": 7,
"input_shapes": {
"input_ids": [
1,
106
],
"attention_mask": [
1,
106
],
"mm_token_type_ids": [
1,
106
],
"pixel_values": [
280,
1536
],
"image_grid_thw": [
1,
3
]
},
"contains_all_shape_colors": true,
"contains_tallest_bar_fact": true,
"passed": true
}
},
"passed": true,
"finished_utc": "2026-08-18T07:45:36+00:00",
"elapsed_seconds": 3688.98,
"regrade": {
"utc": "2026-08-18T07:46:49+00:00",
"source_report_sha256": "3aab1d1b506b00f8261dda7cfd8f48235608c1dafe74c3e3e60e55424bc97798",
"reason": "Compact requested vision answer now accepts the exact '<letter> <height>' form."
}
}