ncky/timecapsulellm-v2-llama-1.2b-gguf overview

{
  "metadata": {},
  "card_data": {
    "base_model": "haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B",
    "language": [
      "en"
    ],
    "library_name": "transformers",
    "license": "mit",
    "datasets": [
      "postgrammar/london-llm-1800"
    ],
    "quantized_by": "ncky",
    "tags": [
      "text-generation-inference",
      "transformers",
      "llama",
      "gguf",
      "historical"
    ],
    "frontmatter": {
      "base_model": "haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B",
      "language": [
        "en"
      ],
      "library_name": "transformers",
      "license": "mit",
      "datasets": [
        "postgrammar/london-llm-1800"
      ],
      "quantized_by": "ncky",
      "tags": [
        "text-generation-inference",
        "transformers",
        "llama",
        "gguf",
        "historical"
      ]
    },
    "hero_image_url": "",
    "summary": "## About static and imatrix-assisted GGUF quants of https://huggingface.co/haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B. Generated with llama.cpp build 8044 (91ea5d67f). IQ4_XS was quantized with an imatrix generated on 19th-century public-domain English text. Note: this model has FFN dimensions (5504) not divisible by 256, so llama.cpp applied fallback quantization to 22 tensors for K/IQ quant types.",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nbase_model: haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B\nlanguage:\n- en\nlibrary_name: transformers\nlicense: mit\ndatasets:\n- postgrammar/london-llm-1800\nquantized_by: ncky\ntags:\n- text-generation-inference\n- transformers\n- llama\n- gguf\n- historical\n---\n## About\n\nstatic and imatrix-assisted GGUF quants of https://huggingface.co/haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B.\n\nGenerated with `llama.cpp` build `8044` (`91ea5d67f`).\n\n`IQ4_XS` was quantized with an imatrix generated on 19th-century public-domain English text.\n\nNote: this model has FFN dimensions (`5504`) not divisible by `256`, so `llama.cpp` applied fallback quantization to 22 tensors for K/IQ quant types.\n\n## Base Model Info (from original model card)\n\nSource: https://huggingface.co/haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B\n\n| Detail | Value |\n| :--- | :--- |\n| Model Architecture | LlamaForCausalLM (decoder-only transformer) |\n| Parameter Count | ~1.22B |\n| Training Type | Trained from scratch (random initialization) |\n| Tokenizer | Custom BPE, vocab size 32,000 |\n| Sequence Length | 2048 |\n| Attention Type | Grouped Query Attention (16 Q heads / 8 KV heads) |\n| Hidden Size | 2048 |\n| Intermediate Size | 5504 |\n| Layers | 22 |\n\nTraining details reported by the source model card:\n- Final training loss: 3.3951\n- Start training loss: 10.7932\n- Training steps: 182,000\n- Epochs: 0.4997\n- Training time: 117h 51m\n- Reported training cost: $340.97 on an H100 SXM (RunPod)\n\n## Usage\n\nIf you are unsure how to use GGUF files, refer to one of [TheBloke's\nREADMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for\nmore details.\n\n## Provided Quants\n\n(sorted by size, not necessarily quality)\n\n| Link | Type | Size/GB | Notes |\n|:-----|:-----|--------:|:------|\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q2_K.gguf) | Q2_K | 0.5 | smallest |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q3_K_S.gguf) | Q3_K_S | 0.6 | low VRAM |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q3_K_M.gguf) | Q3_K_M | 0.6 | balanced low size |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q3_K_L.gguf) | Q3_K_L | 0.6 | better than Q3_K_M |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.IQ4_XS.gguf) | IQ4_XS | 0.6 | imatrix, recommended at this size |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q4_K_S.gguf) | Q4_K_S | 0.7 | fast, recommended |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q4_K_M.gguf) | Q4_K_M | 0.7 | fast, recommended |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q5_K_S.gguf) | Q5_K_S | 0.8 | higher quality |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q5_K_M.gguf) | Q5_K_M | 0.9 | higher quality |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q6_K.gguf) | Q6_K | 1.0 | very good quality |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.Q8_0.gguf) | Q8_0 | 1.2 | fast, best quality |\n| [GGUF](https://huggingface.co/ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF/resolve/main/TimeCapsuleLLM-v2-llama-1.2B.f16.gguf) | f16 | 2.3 | 16 bpw, overkill |\n",
    "related_quantizations": []
  },
  "tags": [
    "transformers",
    "gguf",
    "text-generation-inference",
    "llama",
    "historical",
    "en",
    "dataset:postgrammar/london-llm-1800",
    "base_model:haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B",
    "base_model:quantized:haykgrigorian/TimeCapsuleLLM-v2-llama-1.2B",
    "license:mit",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 1,
  "downloads": 97,
  "gated": false,
  "private": false,
  "last_modified": "2026-02-14T07:40:22.000Z",
  "created_at": "2026-02-14T07:17:09.000Z",
  "pipeline_tag": "",
  "library_name": "transformers"
}

{
  "_id": "699021753883cdc4e0d4d32f",
  "id": "ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF",
  "modelId": "ncky/TimeCapsuleLLM-v2-llama-1.2B-GGUF",
  "sha": "4f0ce2544db72ba0dd8848edb27cee9548076c56",
  "createdAt": "2026-02-14T07:17:09.000Z",
  "lastModified": "2026-02-14T07:40:22.000Z",
  "author": "ncky",
  "downloads": 97,
  "likes": 1,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "transformers",
  "siblings_count": 15
}