metadata
license: apache-2.0
language:
- en
tags:
- text-generation-inference
- transformers
- smolify
- dslm
pipeline_tag: text-generation
inference:
parameters:
temperature: 1
top_p: 0.95
top_k: 64
🤏 smolified-errmind-compiler-explainer-final
Intelligence, Distilled.
This is a Domain Specific Language Model (DSLM) generated by the Smolify Foundry.
It has been synthetically distilled from SOTA reasoning engines into a high-efficiency architecture, optimized for deployment on edge hardware (CPU/NPU) or low-VRAM environments.
📦 Asset Details
- Origin: Smolify Foundry (Job ID:
4ea62cd3) - Architecture: DSLM-Micro (270M Parameter Class)
- Training Method: Proprietary Neural Distillation
- Optimization: 4-bit Quantized / FP16 Mixed
- Dataset: Link to Dataset
🚀 Usage (Inference)
This model is compatible with standard inference backends like vLLM.
# Example: Running your Sovereign Model
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "smolify/smolified-errmind-compiler-explainer-final"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
messages = [
{'role': 'system', 'content': '''You are a precise expert at interpreting C++ compiler errors (GCC/Clang). Given a raw compiler error log, output ONLY valid JSON with keys: summary, why, fix, patch, confidence. No extra text.'''},
{'role': 'user', 'content': '''tmp_gen/err_undefined_variable_84.cpp:1:5: error: 'v84_sdlk' was not declared in this scope 1 | int main(){ v84_sdlk; // force generic error for undefined_variable }'''}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
add_generation_prompt = True,
).removeprefix('<bos>')
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 1000,
temperature = 1, top_p = 0.95, top_k = 64,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
⚖️ License & Ownership
This model weights are a sovereign asset owned by smolify. Generated via Smolify.ai.
