Upload Interpreter-Qwen3-1.7B-ReTok (piece tokenizer, SFT->CPO->GRPO)

README.md

@@ -0,0 +1,96 @@
+---
+license: apache-2.0
+language:
+- zh
+- en
+pipeline_tag: translation
+tags:
+- translation
+- zh-en
+- en-zh
+- piece-tokenizer
+- retok
+- sft
+- cpo
+- grpo
+---
+
+# Interpreter-Qwen3-1.7B-ReTok
+
+A 1.7B Chinese↔English translation model — the **PieceTokenizer (ReTok) A/B sibling** of
+[`Ismantic/Interpreter-Qwen3-1.7B`](https://huggingface.co/Ismantic/Interpreter-Qwen3-1.7B).
+Same SFT→CPO→GRPO pipeline, same data, same losses and rewards; the only differences are
+the **tokenizer** (a custom SentencePiece-style *piece* tokenizer instead of Qwen3's HF BPE)
+and the base checkpoint (a retokenized "phase-2 v18" piece base).
+
+> ⚠️ **This is not a plug-and-play HF model.** The weights are standard, but the tokenizer is
+> a compiled C++ `piece_tokenizer` extension + 5 sidecar files — `AutoTokenizer.from_pretrained`
+> does **not** work. You must load it through the `PieceTokenizerWrapper` from the
+> [Interpreter repo](https://github.com/Ismantic/Interpreter) (`ReTok/lib/`). See **Usage** below.
+
+## Results (WMT23, COMET = `Unbabel/wmt22-comet-da`)
+
+| Stage | zh→en BLEU / COMET | en→zh BLEU / COMET |
+|---|---|---|
+| base (5-shot) | 17.44 / 0.7582 | 38.10 / 0.8255 |
+| SFT | 19.34 / 0.7762 | 40.09 / 0.8392 |
+| CPO | 18.11 / 0.7941 | 31.38 / 0.8480 |
+| **GRPO (this model)** | 18.43 / **0.7967** | 31.79 / **0.8511** |
+
+COMET climbs monotonically from base → GRPO (+0.039 zh→en, +0.026 en→zh). Compared to the
+HF-BPE sibling `Interpreter-Qwen3-1.7B`, the piece tokenizer costs ≈ −0.016 COMET at SFT but
+the gap shrinks to ≈ −0.003 by GRPO — i.e. **CPO/GRPO recover most of the tokenizer-swap tax.**
+
+## Files
+
+Standard weights (`config.json`, `generation_config.json`, `model.safetensors`) plus the 5
+self-contained piece-tokenizer artifacts: `piece.model`, `dict.txt`, `token_mapping.json`,
+`special_tokens_map.json`, `tokenizer_config.json`.
+
+## Usage
+
+Requires the compiled `piece_tokenizer` extension and `PieceTokenizerWrapper` from the
+Interpreter repo. Chat format is `<bos><user>{prompt}<assistant>{response}<eos>` (**not** ChatML),
+`</s>` (id 2) is the stop token.
+
+```python
+# 1) get the loader:  git clone https://github.com/Ismantic/Interpreter
+#    and build/install the piece_tokenizer C++ extension into your venv.
+import sys, torch
+sys.path.insert(0, "Interpreter/ReTok/lib")
+from tokenizer_wrapper import PieceTokenizerWrapper
+from transformers import AutoModelForCausalLM
+
+model_dir = "path/to/Interpreter-Qwen3-1.7B-ReTok"   # this repo, downloaded
+tok   = PieceTokenizerWrapper(model_dir)
+model = AutoModelForCausalLM.from_pretrained(model_dir, torch_dtype=torch.bfloat16).cuda().eval()
+
+def translate(text, direction="zh2en"):
+    if direction == "zh2en":
+        instr = f"Translate the following text from Chinese to English.\nChinese: {text}\nEnglish:"
+    else:
+        instr = f"Translate the following text from English to Chinese.\nEnglish: {text}\nChinese:"
+    ids = tok.apply_chat_template([{"role": "user", "content": instr}],
+                                  tokenize=True, add_generation_prompt=True)
+    out = model.generate(torch.tensor([ids]).cuda(), max_new_tokens=256,
+                         do_sample=False, eos_token_id=tok.eos_token_id)
+    return tok.decode(list(out[0][len(ids):]), skip_special_tokens=True).strip()
+
+print(translate("人工智能正在深刻改变我们的生活方式。", "zh2en"))
+```
+
+vLLM works with `skip_tokenizer_init=True` + `TokensPrompt(prompt_token_ids=...)` — see
+`ReTok/eval/eval_vllm_piece.py` and `ReTok/RUN_BEST_MODEL.md` in the repo for a batch runner.
+
+## Training
+
+- Base: retokenized piece "phase-2 v18_tie" checkpoint (1.7B).
+- SFT: full fine-tune on piece-format translation pairs (ALMA + X-ALMA, ~36.8K).
+- CPO: LoRA preference training (DPO loss + NLL) on ~44K self-generated pairs, then merged.
+- GRPO: full-parameter RL, reward = `wmt22-comet-da` COMET (1.0) + 4-gram repetition penalty (0.3),
+  WMT17–21 source prompts. Reward/hyperparams are identical to the Qwen sibling (A/B contract).
+
+## License & attribution
+
+Apache-2.0. Derived from the Qwen3-1.7B model family and the ALMA / X-ALMA parallel & preference
+corpora + WMT test sets; respect the upstream licenses. Piece tokenizer built for this project.

config.json

@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "dtype": "bfloat16",
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 81899,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 81903
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "bos_token_id": 1,
+  "eos_token_id": [
+    2
+  ],
+  "max_new_tokens": 2048,
+  "pad_token_id": 81899,
+  "transformers_version": "4.57.6"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12dd61b56529a494d820eb9048811706ecec94f1019bce07c749f56c37454c78
+size 3154330384

piece.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9b81cefcaa5d47cd3aa6e653dda0a80f90b7863b3cfff790dfc07c662dda50f
+size 2707732

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "unk_token": "<unk>",
+  "pad_token": "<pad>"
+}

token_mapping.json

@@ -0,0 +1,17 @@
+{
+  "base_vocab_size": 81903,
+  "total_vocab_size": 81903,
+  "special_tokens": {
+    "<pad>": 81899,
+    "<user>": 81900,
+    "<assistant>": 81901,
+    "<system>": 81902
+  },
+  "bos_id": 1,
+  "eos_id": 2,
+  "unk_id": 0,
+  "pad_id": 81899,
+  "user_id": 81900,
+  "assistant_id": 81901,
+  "system_id": 81902
+}

tokenizer_config.json

@@ -0,0 +1,9 @@
+{
+  "model_type": "qwen3",
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "unk_token": "<unk>",
+  "pad_token": "<pad>",
+  "clean_up_tokenization_spaces": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}