Text Generation
Transformers
Safetensors
English
ivme
language-model
transformer
rope
swiglu
muon
from-scratch
tiny
small
decoder-only
custom_code
Instructions to use IvmeLabs/Ivme-Conversate-v2-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use IvmeLabs/Ivme-Conversate-v2-Base with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="IvmeLabs/Ivme-Conversate-v2-Base", trust_remote_code=True)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("IvmeLabs/Ivme-Conversate-v2-Base", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use IvmeLabs/Ivme-Conversate-v2-Base with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "IvmeLabs/Ivme-Conversate-v2-Base" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "IvmeLabs/Ivme-Conversate-v2-Base", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/IvmeLabs/Ivme-Conversate-v2-Base
- SGLang
How to use IvmeLabs/Ivme-Conversate-v2-Base with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "IvmeLabs/Ivme-Conversate-v2-Base" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "IvmeLabs/Ivme-Conversate-v2-Base", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "IvmeLabs/Ivme-Conversate-v2-Base" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "IvmeLabs/Ivme-Conversate-v2-Base", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use IvmeLabs/Ivme-Conversate-v2-Base with Docker Model Runner:
docker model run hf.co/IvmeLabs/Ivme-Conversate-v2-Base
Delete config.py
Browse files
config.py
DELETED
|
@@ -1,29 +0,0 @@
|
|
| 1 |
-
from dataclasses import dataclass
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
@dataclass
|
| 5 |
-
class IvmeConfig:
|
| 6 |
-
"""Ivme-Conversate-v2 (Dense) architecture config.
|
| 7 |
-
|
| 8 |
-
Every field here corresponds to a decision in Section 4 of the design doc.
|
| 9 |
-
Values are chosen to match v1 wherever the doc calls for it, so any quality
|
| 10 |
-
difference between v1 and v2 is attributable to data/training, not size.
|
| 11 |
-
"""
|
| 12 |
-
|
| 13 |
-
vocab_size: int = 16_000 # Section 4.9: 16k tokens, English-only
|
| 14 |
-
hidden_dim: int = 384 # Section 4.2: matches v1
|
| 15 |
-
n_layers: int = 10 # Section 4.3: matches v1
|
| 16 |
-
n_heads: int = 6 # Section 4.4: full attention, no GQA
|
| 17 |
-
context_len: int = 1024 # Section 4.10: matches v1
|
| 18 |
-
ffn_mult: float = 4.0 # SwiGLU hidden expansion (adjusted below for param parity)
|
| 19 |
-
rope_theta: float = 10_000.0 # standard RoPE base frequency
|
| 20 |
-
norm_eps: float = 1e-5 # RMSNorm epsilon
|
| 21 |
-
tie_embeddings: bool = True # Section 4.8
|
| 22 |
-
dropout: float = 0.0 # no dropout at this data:param ratio (heavily overtrained regime)
|
| 23 |
-
|
| 24 |
-
def __post_init__(self):
|
| 25 |
-
assert self.hidden_dim % self.n_heads == 0, "hidden_dim must be divisible by n_heads"
|
| 26 |
-
|
| 27 |
-
@property
|
| 28 |
-
def head_dim(self) -> int:
|
| 29 |
-
return self.hidden_dim // self.n_heads
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|