npm - @bendyline/gilde - Versions diffs - 0.1.20 → 0.1.21 - Mend

@bendyline/gilde 0.1.20 → 0.1.21

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (28) hide show

package/authoring/chat-models/muse-glimmer-30b-q4.json CHANGED Viewed

@@ -26,7 +26,7 @@
   "supportsTools": true,
   "contextWindow": 131072,
   "upstream": "https://huggingface.co/meta-models/Muse-Glimmer-30B",
-  "minGezelVersion": "1.99999",
+  "minGezelVersion": "1.26223",
   "style": {
     "family": "muse",
     "reasoningFormat": "channel",

package/authoring/chat-models/nemotron3.5-lightning-30b-q4.json ADDED Viewed

@@ -0,0 +1,134 @@
+{
+  "id": "nemotron3.5-lightning-30b-q4",
+  "name": "Nemotron 3.5 Lightning (30B-A3B, Q4)",
+  "description": "NVIDIA's Nemotron 3.5 Lightning — a text-only 30B-parameter hybrid Mamba-2/Transformer Mixture-of-Experts model with about 3B active parameters per token, native reasoning and tool calling, and a 256K context window. The llama.cpp source is ggml-org's Q4_K_M build (about 25.4GB); Apple Silicon uses mlx-community's matching 4-bit affine conversion (about 17.8GB).",
+  "tags": [
+    "nvidia",
+    "reasoning",
+    "agentic",
+    "coding",
+    "tools",
+    "long-context",
+    "moe",
+    "hybrid",
+    "mamba",
+    "open-weights",
+    "apple-silicon"
+  ],
+  "category": "reasoning",
+  "maintainer": {
+    "name": "NVIDIA",
+    "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16"
+  },
+  "version": "1.0.0",
+  "updatedAt": "2026-08-11T00:00:00Z",
+  "license": "OpenMDW-1.1",
+  "licenseClass": "open",
+  "licenseShortName": "OpenMDW 1.1",
+  "licenseUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/blob/main/LICENSE",
+  "parameterSize": "30B",
+  "approxSizeBytes": 25430738944,
+  "supportsTools": true,
+  "contextWindow": 262144,
+  "kvBytesPerTokenF16": 6144,
+  "upstream": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16",
+  "style": {
+    "family": "nemotron",
+    "reasoningFormat": "think",
+    "toolCallFormat": "function-call"
+  },
+  "behaviors": [
+    "reasoning.strip-think-tags",
+    "prompt.private-reasoning-guidance",
+    "prompt.tool-cookbook-condensed",
+    "mcp.compact-tool-schemas",
+    "fabrication.detect-past-tense-no-tools",
+    "prompt.retrieval-first",
+    "prompt.workspace-gestalt",
+    "turn.preamble-folding",
+    "turn.ramble-detection",
+    "tools.mlx-grammar"
+  ],
+  "tuning": {
+    "sampling": {
+      "temperature": 1,
+      "topP": 0.95,
+      "maxTokens": 16384
+    },
+    "samplingWhenThinking": {
+      "temperature": 1,
+      "topP": 0.95
+    },
+    "reasoning": {
+      "enableThinking": true,
+      "thinkingBudget": 8192
+    },
+    "profiles": {
+      "thinking-general": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 16384
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 8192
+        }
+      },
+      "thinking-coding": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 32768
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 12288
+        }
+      },
+      "thinking-precise": {
+        "sampling": {
+          "temperature": 0.6,
+          "topP": 0.9,
+          "seed": 0,
+          "maxTokens": 8192
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 6144
+        }
+      },
+      "instruct": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 8192
+        },
+        "reasoning": {
+          "enableThinking": false
+        }
+      },
+      "creative": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 16384
+        },
+        "reasoning": {
+          "enableThinking": false
+        }
+      }
+    }
+  },
+  "llamaCpp": {
+    "huggingfaceRepo": "ggml-org/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-GGUF",
+    "filename": "NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Q4_K_M.gguf",
+    "quantization": "Q4_K_M",
+    "residentBytes": 30516886733
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-4bit",
+    "quantization": "4bit",
+    "residentBytes": 23130356200
+  }
+}

package/authoring/chat-models/nemotron3.5-lightning-30b-q8.json ADDED Viewed

@@ -0,0 +1,134 @@
+{
+  "id": "nemotron3.5-lightning-30b-q8",
+  "name": "Nemotron 3.5 Lightning (30B-A3B, Q8)",
+  "description": "NVIDIA's Nemotron 3.5 Lightning — a text-only 30B-parameter hybrid Mamba-2/Transformer Mixture-of-Experts model with about 3B active parameters per token, native reasoning and tool calling, and a 256K context window. This higher-fidelity variant uses ggml-org's Q8_0 llama.cpp build (about 35GB) or mlx-community's matching 8-bit affine conversion (about 33.6GB).",
+  "tags": [
+    "nvidia",
+    "reasoning",
+    "agentic",
+    "coding",
+    "tools",
+    "long-context",
+    "moe",
+    "hybrid",
+    "mamba",
+    "open-weights",
+    "apple-silicon"
+  ],
+  "category": "reasoning",
+  "maintainer": {
+    "name": "NVIDIA",
+    "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16"
+  },
+  "version": "1.0.0",
+  "updatedAt": "2026-08-11T00:00:00Z",
+  "license": "OpenMDW-1.1",
+  "licenseClass": "open",
+  "licenseShortName": "OpenMDW 1.1",
+  "licenseUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/blob/main/LICENSE",
+  "parameterSize": "30B",
+  "approxSizeBytes": 35004642304,
+  "supportsTools": true,
+  "contextWindow": 262144,
+  "kvBytesPerTokenF16": 6144,
+  "upstream": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16",
+  "style": {
+    "family": "nemotron",
+    "reasoningFormat": "think",
+    "toolCallFormat": "function-call"
+  },
+  "behaviors": [
+    "reasoning.strip-think-tags",
+    "prompt.private-reasoning-guidance",
+    "prompt.tool-cookbook-condensed",
+    "mcp.compact-tool-schemas",
+    "fabrication.detect-past-tense-no-tools",
+    "prompt.retrieval-first",
+    "prompt.workspace-gestalt",
+    "turn.preamble-folding",
+    "turn.ramble-detection",
+    "tools.mlx-grammar"
+  ],
+  "tuning": {
+    "sampling": {
+      "temperature": 1,
+      "topP": 0.95,
+      "maxTokens": 16384
+    },
+    "samplingWhenThinking": {
+      "temperature": 1,
+      "topP": 0.95
+    },
+    "reasoning": {
+      "enableThinking": true,
+      "thinkingBudget": 8192
+    },
+    "profiles": {
+      "thinking-general": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 16384
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 8192
+        }
+      },
+      "thinking-coding": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 32768
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 12288
+        }
+      },
+      "thinking-precise": {
+        "sampling": {
+          "temperature": 0.6,
+          "topP": 0.9,
+          "seed": 0,
+          "maxTokens": 8192
+        },
+        "reasoning": {
+          "enableThinking": true,
+          "thinkingBudget": 6144
+        }
+      },
+      "instruct": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 8192
+        },
+        "reasoning": {
+          "enableThinking": false
+        }
+      },
+      "creative": {
+        "sampling": {
+          "temperature": 1,
+          "topP": 0.95,
+          "maxTokens": 16384
+        },
+        "reasoning": {
+          "enableThinking": false
+        }
+      }
+    }
+  },
+  "llamaCpp": {
+    "huggingfaceRepo": "ggml-org/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-GGUF",
+    "filename": "NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Q8_0.gguf",
+    "quantization": "Q8_0",
+    "residentBytes": 42005570765
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit",
+    "quantization": "8bit",
+    "residentBytes": 43650255848
+  }
+}

package/data/chat-models/de/deepseek-v4-flash-284b-mxfp4/manifest.json CHANGED Viewed

@@ -114,6 +114,8 @@
     "residentBytes": 85899345920,
     "quantization": "MXFP4",
     "cacheExpertsBytes": 68719476736,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-mxfp4/versions/1.0.0/manifest.json CHANGED Viewed

@@ -20,6 +20,8 @@
     "residentBytes": 85899345920,
     "quantization": "MXFP4",
     "cacheExpertsBytes": 68719476736,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q2/manifest.json CHANGED Viewed

@@ -114,6 +114,8 @@
     "residentBytes": 38654705664,
     "quantization": "IQ2_XXS",
     "cacheExpertsBytes": 34359738368,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q2/versions/1.0.0/manifest.json CHANGED Viewed

@@ -20,6 +20,8 @@
     "residentBytes": 38654705664,
     "quantization": "IQ2_XXS",
     "cacheExpertsBytes": 34359738368,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q2/versions/1.1.0/manifest.json CHANGED Viewed

@@ -20,6 +20,8 @@
     "residentBytes": 38654705664,
     "quantization": "IQ2_XXS",
     "cacheExpertsBytes": 34359738368,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q2q4/manifest.json CHANGED Viewed

@@ -115,6 +115,8 @@
     "residentBytes": 45097156608,
     "quantization": "IQ2_XXS/Q4_K",
     "cacheExpertsBytes": 38654705664,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q2q4/versions/1.0.0/manifest.json CHANGED Viewed

@@ -20,6 +20,8 @@
     "residentBytes": 45097156608,
     "quantization": "IQ2_XXS/Q4_K",
     "cacheExpertsBytes": 38654705664,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q4/manifest.json CHANGED Viewed

@@ -311,6 +311,8 @@
     "residentBytes": 85899345920,
     "quantization": "Q4_K",
     "cacheExpertsBytes": 68719476736,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q4/versions/1.0.0/manifest.json CHANGED Viewed

@@ -218,6 +218,8 @@
     "residentBytes": 85899345920,
     "quantization": "Q4_K",
     "cacheExpertsBytes": 68719476736,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/de/deepseek-v4-flash-284b-q4/versions/1.1.0/manifest.json CHANGED Viewed

@@ -218,6 +218,8 @@
     "residentBytes": 85899345920,
     "quantization": "Q4_K",
     "cacheExpertsBytes": 68719476736,
-    "ssdStreaming": true
+    "ssdStreaming": true,
+    "kvBytesPerToken": 8192,
+    "residentCtxTokens": 131072
   }
 }

package/data/chat-models/gl/glm-5.2-754b-q2/manifest.json CHANGED Viewed

@@ -60,6 +60,8 @@
     "residentBytes": 61363217408,
     "cacheExpertsBytes": 34359738368,
     "ssdStreaming": true,
-    "maxLaunchCtx": 65536
+    "maxLaunchCtx": 65536,
+    "kvBytesPerToken": 91136,
+    "residentCtxTokens": 65536
   }
 }

package/data/chat-models/gl/glm-5.2-754b-q2/versions/1.0.0/manifest.json CHANGED Viewed

@@ -13,6 +13,8 @@
     "residentBytes": 61363217408,
     "cacheExpertsBytes": 34359738368,
     "ssdStreaming": true,
-    "maxLaunchCtx": 65536
+    "maxLaunchCtx": 65536,
+    "kvBytesPerToken": 91136,
+    "residentCtxTokens": 65536
   }
 }