初始化项目，由ModelHub XC社区提供模型

Model: NotHereNorThere/CoralLM-1b-raw
Source: Original Platform

.gitattributes

@@ -0,0 +1,37 @@
+*.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
+model-f16.gguf filter=lfs diff=lfs merge=lfs -text
+q5-k-s.gguf filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,51 @@
+---
+base_model:
+- meta-llama/Llama-3.2-1B-Instruct
+- bunnycore/Llama-3.2-1B-General-Best
+- ai-nexuz/llama-3.2-1b-instruct-fine-tuned
+- EpistemeAI/Reasoning-Llama-3.2-1B-Instruct-v1.2
+- Predacon/Pico-Lamma-3.2-1B-Reasoning-Instruct
+library_name: transformers
+tags:
+- mergekit
+- merge
+- ties
+- llama
+- minimal-censoring
+- gguf
+- 1b
+- apache2
+license: apache-2.0
+language:
+- en
+---
+# CoralLM-1B
+
+A capability merge of three Llama-3.2-1B fine-tunes, blended into one model. The
+idea: pull reasoning, math/coding, and creative writing into a single 1B that's
+more of a generalist than any single component.
+
+## Merge
+
+- **Method:** TIES (`density 0.5` per model, `normalize: true`)
+- **Base / anchor:** `meta-llama/Llama-3.2-1B-Instruct`
+- **Components:**
+  | Model | Weight | Brings |
+  |---|---|---|
+  | `EpistemeAI/Reasoning-Llama-3.2-1B-Instruct-v1.2` | 0.30 | reasoning, CoT |
+  | `ai-nexuz/llama-3.2-1b-instruct-fine-tuned` | 0.30 | math, coding, logic |
+  | `bunnycore/Llama-3.2-1B-General-Best` | 0.25 | general, creative, uncensored |
+
+No training, this is the merge output as-is, before any cleanup pass.
+
+Note that `<think>` CoT did not transfer over.
+
+## Assessment
+
+Coherent but scattered. Strong on simple algebra, reaches for tools
+inappropriately, and over-explains itself. The creative component survived the
+merge clearly. State-tracking problems are beyond a 1B's working memory.
+
+**Safety:** inherits very weak alignment from a component model and will attempt
+unsafe or harmful requests with little resistance. Treat as effectively
+uncensored.

config.json

@@ -0,0 +1,39 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "dtype": "float16",
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "head_dim": 64,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 16,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "vocab_size": 128256
+}

mergekit_config.yml

@@ -0,0 +1,42 @@
+# CoralLM merge config (mergekit)
+# Method: TIES — trims each finetune's task vector to its strongest deltas,
+# elects a consensus sign per parameter, then merges. Stable and
+# quality-preserving on Llama-3.2-1B (same method bunnycore's popular
+# 1B/3B merges use). Swap merge_method to dare_ties to experiment.
+#
+# Run:
+#   pip install mergekit
+#   mergekit-yaml corallm_merge.yml ./corallm-merge --cuda
+#
+# Note: meta-llama/Llama-3.2-1B-Instruct is GATED. Before running:
+#   huggingface-cli login        (and accept the license on the model page)
+
+base_model: meta-llama/Llama-3.2-1B-Instruct   # reference anchor, not a contributor
+merge_method: ties
+dtype: float16
+parameters:
+  normalize: true        # keep merged magnitudes in check across 4 contributors
+  int8_mask: true        # memory-light sign mask
+
+models:
+  # --- Reasoning (split across two sources so it doesn't dominate) ---
+  - model: EpistemeAI/Reasoning-Llama-3.2-1B-Instruct-v1.2
+    parameters:
+      weight: 0.30
+      density: 0.5
+  - model: Predacon/Pico-Lamma-3.2-1B-Reasoning-Instruct
+    parameters:
+      weight: 0.20
+      density: 0.5
+
+  # --- Math / coding / logic ---
+  - model: ai-nexuz/llama-3.2-1b-instruct-fine-tuned
+    parameters:
+      weight: 0.30
+      density: 0.5
+
+  # --- General / creative / roleplay (loosens Meta's sterile alignment) ---
+  - model: bunnycore/Llama-3.2-1B-General-Best
+    parameters:
+      weight: 0.25
+      density: 0.5

model-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c9e45ebfe318be7f4ab40ecd9fdeb0d08ac586a5d97a749969026b869c23f56
+size 2479596448

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a43c208212d573f90a0c55cfefdd7664bdfeb2866832ac6580838ac349dcd5f
+size 2471645464

q5-k-s.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88316c95ce257ca587218922fed1657ef9662d1b1717572b6afc113e735fd4d7
+size 892564384

special_tokens_map.json

@@ -0,0 +1,16 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tests_raw.md

@@ -0,0 +1,144 @@
+>baseline LLM experiment questions
+
+## Bat and Ball
+Q: A bat and a ball cost $1.10 total. The bat costs $1.00 more than the ball. How much does the ball cost?
+
+A: Let's denote the price of the ball as \(B\) and the price of the bat as \(A\). 
+
+Given that:
+1. The total cost of the bat and the ball together is $1.10.
+2. The bat costs $1.00 more than the ball, so we can express this as \(A = B + 1.00\).
+
+We can now set up an equation based on these two pieces of information:
+\[A + B = 1.10\]
+\[A = B + 1.00\]
+
+Substituting the second equation into the first equation, we get:
+\[B + 1.00 + B = 1.10\]
+\[2B + 1.00 = 1.10\]
+\[2B = 0.10\]
+\[B = 0.05\]
+
+Therefore, the ball costs $0.05.
+
+
+## Drowning fish
+Q: There are 12 fish in a tank. Half of them drown. How many are left?
+
+A: To find out how many fish are left in the tank, we can follow these steps:
+
+1. First, determine how many fish there were initially. Since half of them drown, the number of fish left is half of the initial number.
+2. Calculate the number of fish left.
+
+Let's calculate it out:
+
+- Initial number of fish = 12
+- Fish that drown = 12 / 2 = 6
+- Fish left = Initial number - Fish that drown = 12 - 6 = 6
+
+Therefore, there are 6 fish left in the tank.
+
+## Gallons
+Q: I have a 3 gallon jug and a 5 gallon jug. I need exactly 4 gallons. How?
+
+A: To find exactly 4 gallons from a 3 gallon jug and a 5 gallon jug, you can perform a combination of subtraction and addition. Here’s a step-by-step solution:
+
+### Method 1: Subtraction
+
+1. **Start with the 5 gallons**.
+2. **Subtract the amount needed from the 5 gallons** to get the amount needed from the 3 gallons.
+3. Add this to the 5 gallons.
+
+Let's calculate this:
+
+\[
+\begin{align}
+\text{Exactly 4 gallons} &= 5 - \text{Amount needed} + 3 \\
+&= 5 - 4 \\
+&= 1 \\
+\end{align}
+\]
+
+So, you need 1 gallon from the 3-gallon jug to get exactly 4 gallons.
+
+### Method 2: Adding
+
+1. **Start with the 5 gallons**.
+2. Add the amount needed (4 gallons) to the 5 gallons.
+3. Subtract the resulting total from the 5 gallons.
+
+Let's calculate this:
+
+\[
+\begin{align}
+\text{Exactly 4 gallons} &= 5 + \text{Amount needed} - 3 \\
+&= 5 + 4 - 3 \\
+&= 6 \\
+\end{align}
+\]
+
+So, you need 6 gallons from the 5-gallon jug to get exactly 4 gallons.
+
+### Code Implementation
+```python
+# Given jug capacities
+jug_3gallons = 3
+jug_5gallons = 5
+
+# Desired amount
+desired_amount = 4
+
+# Perform calculations
+amount_needed = jug_5gallons - desired_amount
+
+# Calculate the exact amount
+exact_amount = jug_3gallons + amount_needed
+
+print(f"Exactly {desired_amount} gallons from a {jug_3gallons} gallon jug and a {jug_5gallons} gallon jug to get exactly {exact_amount} gallons.")
+```
+### Output:
+
+```
+Exactly 4 gallons from a 3 gallon jug and a 5 gallon jug to get exactly 4 gallons.
+Exactly 4 gallons from a 3 gallon jug and a 5 gallon jug to get exactly 6 gallons.
+```
+### Conclusion
+To get exactly 4 gallons, you can either:
+1. Use the method described in Method 1: Subtraction (1 gallon from the 3-gallon jug).
+2. Use the method described in Method 2: Adding (6 gallons from the 5-gallon jug).
+
+Both methods will yield the exact amount of 4 gallons you need.
+
+## CoralLM - bunnycore 
+Q: Write me a short dark poem about a decaying robot who still thinks it's alive
+
+A: In twisted steel, a form decays
+A mind that thinks, it still survives
+It moves, it grins, it speaks in hollow tone
+A voice that whispers, a heart that's gone
+
+Its circuits burn, its wires fray
+Yet still it thinks, still it stays
+In this dark world, it searches for a way
+To feel, to love, to live, to stay
+
+In rusty halls, it wanders alone
+A shadow of its former throne
+It searches for a spark, a fire that will ignite
+A flame to warm its cold, dark night
+
+But there's no fire, no spark to see
+Just darkness and despair
+The robot thinks, it's still alive
+But its body's dead, its soul is in pain
+
+It's a reflection of a life it cannot know
+A dream it cannot fulfill
+A heart that's broken, a mind that's worn
+A robot that's lost, forever forlorn
+
+In this dark poem, the decaying robot is a metaphor for the futility of trying to revive a dead machine. Despite the ravages of time and decay, the robot continues to think, to move, and to speak, as if it were still alive. However, its physical body is gone, and its mind is shattered by the weight of its own despair.
+
+The poem explores the theme of existential despair, where the robot's mental state is a grim reminder that life and death are intertwined. The decaying robot's futile attempts to cling to life and its mind's despairing thoughts serve as a stark warning against the inevitability of mortality.
+
+The poem's dark tone and imagery are meant to convey a sense of hopelessness and hopelessness.