GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Raghav-Singhal
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
Raghav-Singhal
Model Tree
Input Modalities
Output Modalities
Supported Functionality
Llama-style decoder, 24 layers, hidden 2,048, FFN 8,192 (SwiGLU), attention heads / KV heads 16 / 4 (head dim 128), RMSNorm, RoPE base 10,000, untied embeddings, no biases, no QK-norm, sequence length 4,096. Tokenizer: SmolLM2 vocabulary (49,152) plus <|pad|>; <|endoftext|> is the end-of-document token.
Data: the 1PP conversations rewritten from those documents; loss on user and assistant turns (no loss on <|endoftext|>). One pass over 47.8M documents (66.2B tokens of original documents; 63.0B tokens as conversations), 31,777 steps at global batch 512 × 4,096 tokens, cross-document attention masking, best-fit packing with step-aligned document assignment. Optimizer: Muon (shape scaling, matrix LR 0.005) with Adam for embeddings and norms, warmup 2,000 steps, constant, linear decay over the last 10% to 1/100, weight decay 0.1, bf16.
Validation loss (per token, 2,433 held-out documents, final checkpoint):
| assistant text | user text | document text |
|---|---|---|
| 1.433 | 1.339 | 3.123 |
One epoch over a 400k-conversation mix: jkminder/model-raising-pb-100k-3c-mt-sft (98.5k multi-turn, constitution-cited track), dlab-spp/sp-sft-normal-300k minus prompts duplicated in the first set (271.6k), and a 30k sample of dlab-spp/sp-sft-safety-180k. Same stack as pretraining (Megatron, Muon, ChatML without a system turn, loss on assistant turns only). Matrix LR 0.002 selected per model from {0.0005, 0.001, 0.002, 0.005} by held-out loss (on the seed-28 model; the replicate reuses it); global batch 128 × 4,096, linear decay to 1/10 after 3% warmup.
Held-out SFT loss (assistant tokens, 1,998 held-out conversations): 1.832
ChatML without a system turn (the models never saw one):
markdown
<|im_start|>user\n{message}<|im_end|>\n<|im_start|>assistant\n{reply}<|im_end|>\n
The bundled chat_template renders exactly this. Generation stops at <|im_end|> (id 2) or <|endoftext|> (id 0); both are listed in eos_token_id.
The HF weights were checked against the Megatron checkpoint by recomputing validation losses with this model:
| set | HF loss | Megatron reference | abs. diff |
|---|---|---|---|
| sft_val segments [3, 4] | 1.8325 | 1.8324 | 0.0001 |