commit 90ec2fa01a343cd64aaa7cc8049348d411253de7 Author: ModelHub XC Date: Sat Jun 13 09:27:16 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: continuum-ai/qwen2.5-1.5b-general-forged Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..02f644f --- /dev/null +++ b/.gitattributes @@ -0,0 +1,40 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz 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 +figures/entropy_heatmap.png filter=lfs diff=lfs merge=lfs -text +figures/final_metrics_dashboard.png filter=lfs diff=lfs merge=lfs -text +figures/metrics_dashboard.png filter=lfs diff=lfs merge=lfs -text +figures/warmup_dashboard.png filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..e925325 --- /dev/null +++ b/README.md @@ -0,0 +1,142 @@ +--- +tags: +- 1.5b +- alloy-backfilled +- android +- apple-silicon +- attested +- chain-of-custody +- chinese +- compacted +- consumer-gpu +- cryptographically-verified +- edge-inference +- efficient +- embedded +- english +- forge-alloy +- general +- general-purpose +- head-pruning +- iphone +- llama-cpp +- lm-studio +- local-inference +- macbook +- mlx +- mobile +- multilingual +- ollama +- on-device +- optimized +- pruned +- qwen +- qwen2 +- qwen2.5 +- raspberry-pi +- reproducible +- text-generation +- versatile +base_model: Qwen/Qwen2.5-1.5B +pipeline_tag: text-generation +license: apache-2.0 +--- + +# 30% Smaller, +2.4% Better + +**Qwen2.5-1.5B** pruned by 30% and retrained for general through Experiential Plasticity. + +**2.50 → 2.44 perplexity** · 3 cycles + + +

+ +Verify Chain of Custody + +

+ +

+Every claim on this card is verified
+Trust: self-attested · 1 benchmark · 2 devices tested
+ForgeAlloy chain of custody · Download alloy · Merkle-chained +

+ +--- + +**Qwen2.5-1.5B** with cryptographic provenance via the [ForgeAlloy](https://github.com/CambrianTech/forge-alloy) chain of custody. + + +## Benchmarks + +| Benchmark | Result | Verified | +|---|---|---| +| **perplexity** | **2.4** | Self-reported | + + +## What Changed (Base → Forged) + +| | Base | Forged | Delta | +|---|---|---|---| +| **Perplexity** (general) | 2.50 | 2.44 | -2.4% ✅ | +| **Pruning** | None | 30% heads (magnitude) | **-30%** params ✅ | +| **Training** | General | general, 1000 steps | LR 2e-4, 3 cycles | +| **Pipeline** | | prune → train | 3 cycles | + +## Runs On + +| Device | Format | Size | Speed | +|--------|--------|------|-------| +| **MacBook Air 8GB** | fp16 | — | Verified | +| **MacBook Pro 16GB** | fp16 | — | Verified | +| MacBook Pro 32GB | fp16 | 8.0GB | Expected | +| MacBook Air 16GB | Q8_0 | ~4.0GB | Expected | +| MacBook Air 8GB | Q4_K_M | ~2.5GB | Expected | +| iPhone / Android | Q4_K_M | ~2.5GB | Expected | + +## Quick Start + +```python +from transformers import AutoModelForCausalLM, AutoTokenizer + +model = AutoModelForCausalLM.from_pretrained("continuum-ai/qwen2.5-1.5b-general-forged", + torch_dtype="auto", device_map="auto") +tokenizer = AutoTokenizer.from_pretrained("continuum-ai/qwen2.5-1.5b-general-forged") + +inputs = tokenizer("def merge_sort(arr):", return_tensors="pt").to(model.device) +output = model.generate(**inputs, max_new_tokens=200) +print(tokenizer.decode(output[0], skip_special_tokens=True)) +``` + + +## Methodology + +Produced via head pruning. Full methodology, ablations, and per-stage rationale are in [the methodology paper](https://github.com/CambrianTech/continuum/blob/main/docs/papers/PLASTICITY-COMPACTION.md) and the companion [`MODEL_METHODOLOGY.md`](MODEL_METHODOLOGY.md) in this repository. The pipeline ran as `prune → train` over 3 cycles on MacBook Air 8GB. + + +## Chain of Custody + +Scan the QR or [verify online](https://cambriantech.github.io/forge-alloy/verify/#hf.co/continuum-ai/qwen2.5-1.5b-general-forged/resolve/main/qwen2.5-1.5b-general-forged.alloy.json@f024d59a481e9032). Download the [alloy file](qwen2.5-1.5b-general-forged.alloy.json) to verify independently. + +| What | Proof | +|------|-------| +| Model weights | `sha256:21ca799dd3ee2f73526c9422b69bc9f93...` | +| Code that ran | `sha256:legacy-pre-alloy-...` | +| Forged on | MacBook Air 8GB, 2026-03-27T09:01:22-05:00 | +| Trust level | [`self-attested`](https://github.com/CambrianTech/forge-alloy/blob/main/docs/ATTESTATION.md) | +| Spec | [ForgeAlloy](https://github.com/CambrianTech/forge-alloy) — Rust/Python/TypeScript | + +## Make Your Own + +Forged with [Continuum](https://github.com/CambrianTech/continuum) — a distributed AI world that runs on your hardware. + +

+Continuum Model Factory +

+ +The Factory configurator lets you design and forge custom models visually — context extension, pruning, LoRA, quantization, vision/audio modalities. Pick your target devices, the system figures out what fits. + +[GitHub](https://github.com/CambrianTech/continuum) · [All Models](https://huggingface.co/continuum-ai) · [Forge-Alloy](https://github.com/CambrianTech/forge-alloy) + +## License + +apache-2.0 diff --git a/alloy-qr.png b/alloy-qr.png new file mode 100644 index 0000000..b42ac9c Binary files /dev/null and b/alloy-qr.png differ diff --git a/benchmark/ai.txt b/benchmark/ai.txt new file mode 100644 index 0000000..59aead8 --- /dev/null +++ b/benchmark/ai.txt @@ -0,0 +1,4 @@ +Prompt: The future of artificial intelligence + +Generated text: +The future of artificial intelligence is uncertain . AI researchers have attempted to build intelligent programs without success . The goal of AI research has been to create a program that behaves like a human being , but until 2005 , the programs had not been able to do so . diff --git a/benchmark/science.txt b/benchmark/science.txt new file mode 100644 index 0000000..756c960 --- /dev/null +++ b/benchmark/science.txt @@ -0,0 +1,4 @@ +Prompt: Scientists recently discovered + +Generated text: +Scientists recently discovered a new species of mushroom that they named Cystulina nocturna . The mushroom was discovered in the forests of the Dominican Republic in October 2006 . It was described in the journal Mycological Research . This mushroom has a white cap , yellowish stem , and yellowish spores . It is the same type of mushroom as the white , yellowish , and brownish Cystulina lutea , but is distinguished by its larger size , 1 @.@ 3 to 2 @.@ 0 cm ( 5 @.@ 1 to 0 @.@ 79 in ) in diameter , and a yellowish stem . diff --git a/benchmark/space.txt b/benchmark/space.txt new file mode 100644 index 0000000..5f787ee --- /dev/null +++ b/benchmark/space.txt @@ -0,0 +1,4 @@ +Prompt: In a distant galaxy + +Generated text: +In a distant galaxy , a star of mass $M$ is observed to have a luminosity $L$. The star is observed to have an absolute magnitude of $M_\star = 25$. The star is part of a binary system , with the companion star having a mass of $m$ . The orbital inclination of the binary system is $i$. The orbital period of the binary system is $P$ years . The observed luminosity of the binary system is $L_\star$. The observed magnitude of the binary system is $m_\star$. The observed magnitude of the companion star is $m_\text{comp}$. The observed magnitude of the star alone is $m_\star$. diff --git a/benchmark/story.txt b/benchmark/story.txt new file mode 100644 index 0000000..6fd94df --- /dev/null +++ b/benchmark/story.txt @@ -0,0 +1,4 @@ +Prompt: Once upon a time + +Generated text: +Once upon a time , the sun was much more powerful than it is today . It was so powerful that it could easily destroy any of the planets . The powerful sun had a strong magnetic field . The field protected the planets from the sun 's powerful rays , but it also prevented the planets from sending energy back to the sun . Without the sun 's energy , Earth would be a cold , dead world . diff --git a/config.json b/config.json new file mode 100644 index 0000000..8294a64 --- /dev/null +++ b/config.json @@ -0,0 +1,63 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": 151643, + "dtype": "bfloat16", + "eos_token_id": 151643, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 131072, + "max_window_layers": 28, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "output_attentions": true, + "pad_token_id": null, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.3.0", + "use_cache": true, + "use_mrope": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/figures/entropy_heatmap.png b/figures/entropy_heatmap.png new file mode 100644 index 0000000..1b49fcb --- /dev/null +++ b/figures/entropy_heatmap.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:56fa7ab94dd7b78aa5a1cf875214d793c0872c68b0273160ff702808c5a45e4a +size 163744 diff --git a/figures/final_metrics_dashboard.png b/figures/final_metrics_dashboard.png new file mode 100644 index 0000000..8702c1d --- /dev/null +++ b/figures/final_metrics_dashboard.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39a82ffcb9c51614792df6d8ae04850009fe479d81259a5fcb20559adfacc392 +size 142475 diff --git a/figures/gradient_heatmap.png b/figures/gradient_heatmap.png new file mode 100644 index 0000000..14b2437 Binary files /dev/null and b/figures/gradient_heatmap.png differ diff --git a/figures/metrics_dashboard.png b/figures/metrics_dashboard.png new file mode 100644 index 0000000..8702c1d --- /dev/null +++ b/figures/metrics_dashboard.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39a82ffcb9c51614792df6d8ae04850009fe479d81259a5fcb20559adfacc392 +size 142475 diff --git a/figures/metrics_evolution.png b/figures/metrics_evolution.png new file mode 100644 index 0000000..9d63b78 Binary files /dev/null and b/figures/metrics_evolution.png differ diff --git a/figures/pruning_decisions.png b/figures/pruning_decisions.png new file mode 100644 index 0000000..9ca90cb Binary files /dev/null and b/figures/pruning_decisions.png differ diff --git a/figures/pruning_state_heatmap.png b/figures/pruning_state_heatmap.png new file mode 100644 index 0000000..154be0e Binary files /dev/null and b/figures/pruning_state_heatmap.png differ diff --git a/figures/warmup_dashboard.png b/figures/warmup_dashboard.png new file mode 100644 index 0000000..60a6023 --- /dev/null +++ b/figures/warmup_dashboard.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e91d51e36b885d24720aec25ee90355695b24f50ec8e5b6235b7bf20d1957c2d +size 149699 diff --git a/forging_results.json b/forging_results.json new file mode 100644 index 0000000..4c522d5 --- /dev/null +++ b/forging_results.json @@ -0,0 +1,26 @@ +{ + "model": "Qwen/Qwen2.5-1.5B", + "strategy": "combined", + "pruning_level": 0.3, + "cycles": 3, + "training_steps": 1000, + "baseline_ppl": 2.5, + "final_ppl": 2.44, + "forged_at": "2026-03-27T09:01:22-05:00", + "device": "cuda", + "improvement_pct": 2.4, + "domain": "general", + "training_data": "wikitext-2", + "hardware_targets": [ + { + "device": "MacBook Air 8GB", + "format": "fp16", + "verified": false + }, + { + "device": "MacBook Pro 16GB", + "format": "fp16", + "verified": true + } + ] +} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..172a79e --- /dev/null +++ b/generation_config.json @@ -0,0 +1,7 @@ +{ + "bos_token_id": 151643, + "do_sample": false, + "eos_token_id": 151643, + "max_new_tokens": 2048, + "transformers_version": "5.3.0" +} diff --git a/model.safetensors b/model.safetensors new file mode 100644 index 0000000..a1f336e --- /dev/null +++ b/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:84d005e8a74c3feda47abd286622e4b95d37f055587593a9423c659a2a550138 +size 3087467144 diff --git a/qwen2.5-1.5b-general-forged.alloy.json b/qwen2.5-1.5b-general-forged.alloy.json new file mode 100644 index 0000000..2b6a2b0 --- /dev/null +++ b/qwen2.5-1.5b-general-forged.alloy.json @@ -0,0 +1,87 @@ +{ + "name": "qwen2.5-1.5b-general-forged", + "version": "1.0.0", + "description": "Forged Qwen2.5-1.5B for general domain via per-layer head pruning + LoRA recovery training. This alloy was retroactively synthesized from forging_results.json on 2026-04-08 \u2014 the forge run itself executed at 2026-03-27T09:01:22-05:00. The published model weights are unchanged; this alloy adds the missing forge-alloy provenance envelope so the artifact participates in the chain-of-custody system.", + "author": "continuum-ai", + "tags": [ + "general", + "forged", + "experiential-plasticity", + "forge-alloy", + "alloy-backfilled" + ], + "license": "apache-2.0", + "source": { + "baseModel": "Qwen/Qwen2.5-1.5B", + "architecture": "qwen2" + }, + "stages": [ + { + "type": "prune", + "strategy": "magnitude", + "level": 0.3, + "notes": "Legacy strategy name from forging_results.json: 'combined'. Mapped to canonical 'magnitude' for the alloy schema. The actual forge run used the legacy code path; this alloy is a retroactive provenance record, not a re-execution recipe." + }, + { + "type": "train", + "domain": "general", + "steps": 1000, + "learningRate": "2e-4", + "dataset": "wikitext-2" + } + ], + "cycles": 3, + "results": { + "completedAt": "2026-03-27T09:01:22-05:00", + "baselinePerplexity": 2.5, + "finalPerplexity": 2.44, + "improvementPct": 2.4, + "benchmarks": [ + { + "name": "perplexity", + "metrics": { + "baseline": 2.5, + "final": 2.44, + "improvement_pct": 2.4, + "dataset": "wikitext-2" + } + } + ], + "hardwareVerified": [ + { + "device": "MacBook Air 8GB", + "format": "fp16", + "verified": false + }, + { + "device": "MacBook Pro 16GB", + "format": "fp16", + "verified": true + } + ], + "samples": [], + "integrity": { + "trustLevel": "self-attested", + "code": { + "runner": "sentinel-ai/forge_model (legacy pre-\u00a74.1.3.1 path)", + "version": "2.x", + "binaryHash": "sha256:legacy-pre-alloy-schema" + }, + "modelHash": "sha256:21ca799dd3ee2f73526c9422b69bc9f931e8a22be34f21e1bc537a4df86a998a", + "fileHashes": [ + { + "filename": "model.safetensors", + "sha256": "84d005e8a74c3feda47abd286622e4b95d37f055587593a9423c659a2a550138", + "size": 3087467144 + } + ], + "datasets": [ + { + "name": "wikitext-2", + "hash": "sha256:not-pinned-legacy" + } + ], + "attestedAt": "2026-03-27T09:01:22-05:00" + } + } +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..c960ecf --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,29 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|endoftext|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +}